How to Plan a Predictive Maintenance Pilot Project

A predictive maintenance pilot should answer a business question. Can better machine data help the team find problems earlier and make better maintenance decisions? A pilot is not simply a small sensor order. It is a test of the entire process: selecting assets, collecting usable data, reviewing it and acting on what the team learns.

The first project does not need to cover a whole plant. In fact, a smaller and well-designed deployment may give clearer answers. Start with machines that have a real consequence of failure. Define success before installing the hardware. Then make the results easy for operations, maintenance and management to review.

Sensoteq offers a practical combination for this approach. Kappa X can monitor suitable rotating equipment. Tau E can add motion and condition data for vibrating or oscillating assets. Analytix brings alarms, trends and diagnostic information into one platform. Tomarok can help translate that portfolio into a pilot scope built around the machines and site conditions.

DEPLOYMENT GUIDESensoteq

Nine stepsAsset selectionDelivery checklist

Request a Sensoteq Pilot Deployment Plan

Your process, equipment or measurement requirement.

01 METHOD

Nine Steps from Question to Decision

Define what to prove, select assets, install well, review honestly, then decide whether to scale.

Request a pilot plan

Decide What the Pilot Must Prove

A vague goal such as “improve reliability” is difficult to measure. Pick a more specific question. The plant may want to detect a known pump problem earlier. It may want better insight into gearbox bearings. Or it may need to compare the motion of vibrating screens.

The pilot can also test practical questions. Does the wireless network reach the assets? Can the maintenance team work with the alerts? Are the measurements useful under normal operating conditions? These are valid outcomes, even if no failure occurs during the pilot.

Choose a Small, Relevant Group of Assets

Select assets with enough operating hours to generate useful data and a clear reason to monitor them. A machine that is rarely used may not show much within a short evaluation. On the other hand, using only machines that are already close to failure can give a distorted view of the program.

See the asset selection factors below. A pilot can mix a few machine types if there is a clear reason. But avoid turning it into an uncontrolled trial of every asset in the plant.

Define the Measurement Questions

For a motor-pump train, the key questions may relate to bearing vibration, alignment or hydraulic instability. For a gearbox, the team may need high-frequency trends and gear-related patterns. For a vibrating screen, stroke, angle and structure comparison may be the priority.

These questions determine the equipment. Do not order sensors before deciding what needs to be measured. They also influence where the sensors will be installed and what operating information should be recorded.

Survey Mounting Points and Wireless Coverage

A good mounting point is safe, rigid and relevant to the target component. It should also remain accessible for inspection and maintenance. Review guards, hot surfaces, washdown areas, vibration levels and any hazardous-area classification.

Gateway placement is another part of the design. Steelwork and site geometry can block signals. Sensoteq recommends checking line of sight and communication at proposed positions. Its gateways can connect through Wi-Fi, cellular or Ethernet depending on the setup. The plant should also check power and IT access before deployment.

If the site is large, it may need more than one gateway. Published device-capacity figures are not a substitute for an on-site communication survey.

Record the Baseline

Before interpreting alarms, capture what normal looks like. Record operating speed, load where available, machine condition and known defects. Note maintenance work completed just before installation. A baseline from an unhealthy machine should not automatically become the definition of “good”.

The monitoring period should cover the machine’s relevant operating conditions. A fixed number of weeks is not appropriate for every site. A continuous-duty pump may provide useful patterns quickly. A seasonal or intermittent asset may require longer.

Set Practical Alarms and Ownership

Alarms should have meaning. A warning may prompt a review of trends or a field inspection. A critical alert may require urgent investigation under the site procedure. The team should agree on escalation rules before the first notification arrives.

Assign a clear owner for each activity. Who reviews Analytix? Who checks the waveform? Who talks to production? Who creates the work order? Who documents the result? Without this workflow, sensor data can become another unattended dashboard.

Review What the Data Actually Shows

During the pilot, look at trends, false alarms, operating context and data quality. The goal is not to celebrate the number of notifications. It is to learn whether the alerts lead to good decisions.

A developing fault may be detected. That is valuable, but not guaranteed. The absence of a failure does not prove the pilot failed. A stable baseline, reliable communication and faster investigation can also be useful outcomes. Each should be judged against the goals set at the start.

Evaluate the Business Case

Review the full cost of the pilot. Include sensors, gateways, software, installation, staff time and specialist analysis. Compare that investment with the expected benefit on the selected machines. Relevant benefits might include earlier intervention, reduced manual visits, fewer emergency repairs and better maintenance planning.

Use plant-specific values. Avoid applying an industry-wide savings percentage to every asset. The predictive vs reactive maintenance cost comparison explains which cost categories to include when building the business case.

Decide Whether and How to Scale

At the end, produce a short decision report. What worked? What did not? Which assets provided useful data? Were sensor positions and gateways reliable? Could the team act on the findings? What should change in a second phase?

Then build the expansion plan from the answers. The plant may add more pump trains, extend into gearboxes or include vibrating screens. It may also need better internal alert ownership before adding equipment. Buying more sensors is not the only measure of pilot success.

02 STEP 2 DETAIL

Asset Selection Factors

ASSET SELECTION FACTOR What to look for Why it matters
Criticality Production, safety or quality impact Makes the result relevant to management
Failure history Known repeat problems or costly interventions Defines a useful monitoring question
Operating pattern Enough run time and stable reference conditions Supports meaningful trends
Accessibility Safe installation and realistic inspections Reduces deployment delays
Data availability Speed, process history and maintenance records Helps explain vibration changes
Connectivity Practical gateway position and backhaul Tests whether the system can scale

03 CHECKLIST

Pilot Delivery Checklist

  • Pilot owner and site stakeholders identified
  • Monitoring problem and acceptance criteria agreed
  • Critical asset list and failure history reviewed
  • Sensor type and measuring points approved
  • Area classification and safe mounting confirmed
  • Gateway coverage and internet connection tested
  • Baseline and operating data collected
  • Alert thresholds and response workflow defined
  • Review cadence and evidence log agreed
  • Final decision report and expansion options prepared

Request a Sensoteq Pilot Deployment Plan

Tell Tomarok which assets matter, what has gone wrong before and what you want the pilot to prove. An equipment list and a site layout are useful starting points. We can review a practical sensor, gateway and software scope, along with the information needed for installation and follow-up. If the project needs a commercial estimate, see our wireless condition monitoring system cost guide before requesting a quotation.

NEWSLETTER

Project news, engineering insights and partner updates — a few times a year, no noise.

Download our company brochure

Enter your work email to get instant access to the Tomarok company brochure. We will also email you the download link.

Work email *

We use your email to send you the brochure. Read our Privacy Policy.