Predicting the journey to improved reliability

As predictive maintenance (PdM) moves from a concept into a practical solution, it is transforming manufacturing operations.

By Gaurav Verma February 13, 2023
Courtesy: Cincinnati Incorporated/Steve Rourke, CFE Media and Technology

Predictive maintenance (PdM) is maintenance that monitors the performance and condition of equipment during normal operation to reduce the likelihood of failures.

The difference between predictive and preventive maintenance – which relies on work being undertaken on an asset before a problem is noticed – lies in the methods used, the amount of lead-time you have for a task, and the precision of scheduling. PdM uses condition-monitoring tools and techniques and asset information to track equipment performance to anticipate failure before it happens. Ideally, PdM keeps maintenance frequency low while reducing time spent on unplanned maintenance and unnecessary preventive maintenance.

PdM relies on tools such as smart sensors which are able to detect a change in the way assets are operating, such as the vibration of a part at higher-than-normal-speeds. The sensors connect with maintenance software, like a computerized maintenance management system (CMMS), which is able to schedule timely maintenance. The software can also notify technicians of the newly scheduled task via mobile devices.

Because predictive maintenance offers a window for proactive maintenance tasks, it can help minimize the time involved in equipment maintenance, the production hours lost to maintenance, and the cost of spare parts and supplies.

According to Allied Market Research, the manufacturing predictive analytics market size valued at $535 million in 2018, is projected to reach $2.5 billion by 2026. This demonstrates the value placed on this type of maintenance – which relies on technology and five critical organizational factors – people, data, processes, tools and parts, and equipment. These factors, combined with technology, are known as the six pillars of a strong predictive maintenance program.

Each of these six pillars can be developed and used to build a predictive maintenance program that will last.

1. People: It doesn’t matter if your predictive maintenance plan looks good on paper if you don’t have buy-in from the people who are doing the work. Each predictive maintenance pillar in a program needs people to build and maintain it. Data needs interpreting. Technology needs designing and managing so everyone within an organization must understand how PdM works, why it’s important and what they can do to make it successful. Getting people onboard with the changes that come with predictive maintenance is essential, but not always easy. It is important to get buy-in from the maintenance team and create a culture of success at the facility.

2. Data: A predictive maintenance program needs information to be successful. Without the data, you can’t predict anything. If you don’t have a baseline about what’s normal for a pump or a conveyor, for example, you can’t identify or predict anomalies. Importantly, there is a need for data quality. The key is to have the right information coming from the plant floor. Data is the link between current asset performance and the future state of the asset, which is why everything – from throughput to failure modes – must be constantly updated. These numbers must also be accurate. If they are different from system to system, it will derail the program.

3. Processes: People processes involve the way the maintenance team goes about their work. They outline how staff interact with machines, data, each other and everything else. It is important to understand who is responsible for what, how frequently data and tasks are reviewed, how you communicate, and how you plan, escalate and complete tasks. When it comes to equipment processes, it is crucial to know what processes the equipment completes, how to capture asset data, and how the data maps to future performance. The processes are the way you work – how the maintenance team plans and does the things it needs to do every day to be successful. An effective predictive maintenance program helps make the entire operation predictable, maximizing everything from working hours to asset performance.

4. Tools and parts: Predictive maintenance isn’t new. The difference between 20 or 30 years ago and now is that we have the tools and understanding of parts to do it better with fewer costs. Tools are the instruments used to measure the condition of assets, like infrared cameras, and the tools used to inspect or repair equipment. Parts are the different components of the equipment.

5. Equipment: It is important to know which equipment allows you to anticipate failure when designing a predictive maintenance program. The assets that fit into a predictive maintenance program are the ones that provide good condition data with enough lead-time to catch problems before total failure. Applying predictive maintenance to the most critical assets with the most observable failure modes is advisable as a first step because of the time and money involved in building a PdM program.

6. Technology: Technology helps manage, facilitate and optimize the other pillars of predictive maintenance. Critical to success is knowing what products are run and when, the cost of all the activities, and when maintenance was last performed among other factors.There are lots of different technologies that can be used to manage a predictive maintenance program – from ERPs to MES systems and CMMS software. A predictive maintenance program will be able to solve every maintenance problem but there are many benefits to having one, like a more reliable operation that allows everyone within an organization to grow and become more efficient. Taking advantage of those benefits involves building on key maintenance fundamentals. When those fundamentals are strong, you’ll have a winning strategy.

– This originally appeared on Control Engineering Europe’s website.


Author Bio: Gaurav Verma is Senior Manager, Software Solutions Marketing at Rockwell Automation.