How to Create a Predictive Maintenance Program

Creating a maintenance program that leverages predictive maintenance (PdM) technologies boosts equipment reliability and strengthens maintenance teams. It allows skilled technicians to focus on assets that are important to the organization while boosting overall team results.

The keys to success for establishing a strong predictive maintenance program are organization and involvement of a broad membership of internal team members. Buy in from the entire team is crucial for successful adoption.

In this article, we’ll walk you through the eights steps that are necessary to achieve success with your first predictive maintenance program.

Step 1: Build a PdM team

Including members of a PdM team outside of the maintenance department ensures success. A broad membership gives your efforts more power by building advocates who understand the goals of predictive maintenance. (This step is crucial which is why we’re emphasizing it here as well as in the introduction to this article.)

PdM team members could include key individuals like production or customer service managers. Folks in the warehouse can contribute, too. They’ll certainly have a lot of opinion about how to improve the PdM program when a freezer unit fails or when a palletizer doesn’t work and they must stack or wrap pallets by hand.

When different departments understand the PdM process, it gives everyone confidence in the process and the work to follow.

Step 2: Select a PdM champion

A champion leads the team building the predictive maintenance program. The champion of the PdM team should be familiar with the technical details of asset maintenance and understand how different departments work together.

The goal of the champion and PdM team is to:

  • Select equipment for inclusion in the PdM program, ranked by priority level
  • Determine resources that are available for the PdM program
  • Group the equipment by the PdM technology that suits it
  • Purchase PdM equipment, services, and training for the initial effort. (Your team may find it useful to contract a PdM vendor for initial readings while teaching team members on the use of the equipment and reports.)
  • Monitor the PdM process and review results to determine what adjustments are necessary

Step 3: Define goals and resources

Your computerized maintenance management system (CMMS) can provide key information about assets and their components. It may also include important items like the priority of certain assets.

The PdM team will benefit by reviewing asset priority. Understanding asset priority helps the team determine the order of work needed for PdM implementation. Not all team members will agree with the rankings. However, the rankings should agree with the process previously developed by asset leaders or the criteria established by the team.

Establishing what resources are available shows the importance of asset ranking. The primary resource constraint is usually money. Money is necessary to obtain the PdM equipment and pay the employees or contractors to use it.

Most PdM technologies take a similar level of work. A good rule of thumb is approximately 30 minutes per asset to complete most PdM assessments. However, assessment time decreases as the organization gets better at the process. Team members become more efficient and some equipment is even dropped or added for other equipment.

Step 4: Select a method for data acquisition

Using a PdM involves getting the data collector device ready, gathering the data, downloading the data, and reviewing the data. These activities take 30 minutes per each asset at the beginning of the program. That is why most portable PdM collection is done on what is called a “Route.” The PdM technician prepares a collector for the equipment that is done on a Route: gather the data, download the data, and then find the statistical outliers for follow up.

Wired systems do not take as much labor to collect and analyze data. But wiring or configuring these systems take more time in the early phases of the program. Many assets do not require this sort of dedicated communication.

Wireless systems offer an alternative between wired and portable collectors. These are often referred to as IoT devices. They take scheduled or continuous readings and send data to your system via WiFi.

Step 5: Define frequencies for readings (and meet them)

A key element of PdM is comparing current readings with previous readings. The regularity of the readings is what helps identify potentially failing components. It also establishes resource limitations. Many assets may only need tested once every 30 days.

A full-time employee dedicated to PdM can adequately monitor 320 assets per month.

(16 Assets Per 8-Hour Day x 5 Days x 4 Weeks)

To help achieve the recommended frequency, resources can be stretched. For instance, you can use a non-technical employee to collect data. Using non-technical employees lower your PdM costs and free up trained technicians to respond to findings.

Step 6: Select technology, tools, and vendors

Purchasing PdM equipment or services should match the equipment that best fits its proven abilities. UpKeep’s Learning Center features a description of the PdM options that are available to teams. Organizers of a PdM effort have a responsibility to identify which technology works best with its assets.

It is important that PdM organizers understand that some tools are better at finding issues than others. Some PdM vendors will claim a wide variety of things their equipment can accomplish, but not every tool is created equal. Most PdM tools can perform a wide variety of tasks, yet some are better at specific tasks than others.

For example, vibration analysis (VA) is best for rotating equipment and VA vendors have powerful software and equipment to quickly identify issues with rotating equipment. Ultrasonic analysis (UA) systems can also find a noisy failing rotating equipment component. The problem is that it doesn’t have some of the features of a VA system.

Step 7: Use multiple PdM tools for a single asset

A problem can be solved by using two PdM technologies simultaneously on a single asset.

For example, a whey processing plant was experiencing equipment failures with one of its large centrifugal units. The units spin at a very high speed and their stainless steel skin makes it difficult to use portable PdM technology and fixed sensors were not installed.

The VA analyzer determined that the issue was likely a bearing, but the location of the bearing was difficult to discover because of the skin. Taking the centrifuge offline and putting multiple expensive parts on it was a difficult choice.

The whey processor used their UA tool to determine the location of where the failing bearing was probably located. UA could accomplish this because sound emissions are directional. Using the UA tool around the machine found the location where the noise was highest and those parts were replaced. The centrifuge was successfully brought back online following the suspect bearings replacement.

The processor also attempted to use their infrared (IR) resources on the centrifuge but it would not work. The camera only showed the image of the person using the sensor as it picked up the reflection from the centrifuge’s stainless steel skin. However, had there been an electrical concern, the IR unit would have quickly depicted the issue before others.

Step 8: Monitor performance and commit to excellence

Hold PdM teams accountable for their performance. They should consistently review findings from the PdM process and celebrate their victories. Great teams look for the lessons in missed results.

The PdM program can grow as it proves its effectiveness by reducing emergency and unplanned failures. The power of the committed team will make this happen.