Why predictive maintenance is key in manufacturing

Historically, manufacturers had three options for managing the lifecycle of equipment:   

• run to fail   
• schedule preventive maintenance   
• guess when an intervention could prevent a failure

With vast streams of sensor-generated data and advanced analytical tools, however, manufacturers are increasingly able to rely on predictive maintenance to ensure maximum return on investment from their equipment.

Manufacturers rely on a broad array of mechanical and electrical equipment, from the shop floor to the loading dock to the distribution fleet to a customer site. Taking equipment offline for preventive maintenance can be just as disruptive as a temporary failure.

According to analysts with the consulting firm Deloitte:

“Traditionally, this dilemma forced most maintenance organizations into a tradeoff situation where they had to choose between maximizing the useful life of a part at the risk of machine downtime (run-to-failure) or attempting to maximize uptime through early replacement of potentially good parts (time-based preventive maintenance), which has been demonstrated to be ineffective for most equipment components.”

Preventive maintenance

Preventive maintenance (PM) is based on history and experience. Observations of the impact of time and usage on manufacturing components are recorded and used to create schedules for performing crucial maintenance designed to prolong the life of those components.

However, as Mark Lamendola, former technical editor of EC&M, writes: “Well-written maintenance procedures don’t cover every contingency. Instead, they are designed to ensure the same maintenance steps are performed each time.”

Predictive maintenance, on the other hand, “focuses on the continuous monitoring of a machine's health under normal working conditions — without process interruptions — to detect subtle changes that generally aren't detectable during typical inspection processes,” Nicole Pontius writes in Business.com. “These subtle warning signs then become the triggers that indicate an impending problem, enabling operators and maintenance providers to plan for repairs and downtime, order necessary parts, and take other preparatory action to minimize the eventual disruption in processes.”

That sounds akin to the demands for agility and continuous improvement that have been driving IT for the past decade and should be no surprise. As digital technologies transform manufacturing, the worlds of IT and operations management are increasingly intertwined. Advanced analytical systems, often driven by machine learning and artificial intelligence, are leveraging Industrial Internet of Things (IIoT) sensors to radically transform manufacturing operations from the realms of mechanical and electrical engineering to digital.

Industry 4.0 and predictive maintenance

The advent of digital manufacturing, sometimes referred to as Industry 4.0, “introduces what has been called the ‘smart factory,’ in which cyber-physical systems monitor the physical processes of the factory and make decentralized decisions,” writes Forbes contributor Bernard Marr.

Not only does this make for better products and enhanced productivity, but it also heralds a new era of predictive maintenance.

Predictive maintenance solutions to support digital optimization represent the top investment category, ahead of investment in IoT and mobile app development. So say 42% of respondents in IT and operations management positions in an IDG Research survey with AT&T. Respondents were almost evenly split among IT and operations-related roles within the manufacturing, CPG, and transportation/fleet industries.

“Predictive maintenance has the potential to add significant value to production processes by increasing efficiency and reducing unplanned and redundant costs,” writes Isaac Maw of Engineering.com. “Additionally, the capability for better analysis of IIoT data makes IIoT devices more valuable, as more and more uses for the data are discovered.”

Gathering data

Data is obviously key to digital manufacturing operations. “Process industries generate enormous volumes of data, but many have failed to make use of this mountain of potential intelligence,” according to McKinsey analysts.

The McKinsey analysts explain:

“Predictive maintenance systems gather historical data (structured and unstructured, machine- and non-machine-based) to generate insights that can’t be observed with conventional techniques. Using advanced analytics, companies can determine the circumstances that tend to cause a machine to break and monitor input parameters so they can intervene before breakage happens — or be ready to replace it when it does — thus minimizing downtime. Predictive maintenance typically reduces machine downtime by 30% to 50% and increases machine life by 20% to 40%.”

Implementing predictive maintenance must go beyond collecting data. “Organizations have to define which data should be collected and how to analyze it automatically, and decide how and when to follow findings with actions,” Ernst & Young analysts advise. “So predictive maintenance requires technical prerequisites as well as changes to the organization, processes and personnel skills.”

A key technology prerequisite is the connectivity to capture, integrate and send data from IoT devices and machines to analytics systems in the cloud and on edge systems. The power and scalability of the cloud make it possible to analyze trends and create predictive maintenance models. Network edge devices are positioned to provide the microsecond decisions that can prevent a device from overheating or otherwise malfunctioning.

“Unlike other technologies that revolve around one predominant architecture, device type, or connection method, IoT is, at its core, an assembly of disparate technologies,” AT&T analysts observe. “A connected machine does not become ‘smart’ from a single sensor, or modem, or network, or application alone. It is a combination of all of these pieces coming together that creates added intelligence.”

Predictive vs. preventive maintenance is not a binary choice.

“Applying a predictive maintenance solution to every asset in a facility just isn’t cost-efficient,” writes Dan Davis in a recent blog post. The reality is that most manufacturers will find they’re striving toward a “properly balanced predictive/preventive maintenance strategy,” Davis writes.

Implementing predictive maintenance requires integrating things such as:   

• device hardware   
• connectivity   
• data management   
• applications   
• analytics   
• security

The constant flow of data requires automation platforms that can facilitate and orchestrate key interactions among each of these layers as well as with other back-end systems in a business.

Visit our manufacturing page for more information on the digital factory and industry-specific solutions and trends.