Optimising predictive maintenance

The benefits of predictive maintenance (PdM) are well established. Improved productivity, lowered maintenance cost and a reduction in unplanned downtime of up to 50 per cent are just a few of the advantages offered. But how do you achieve a truly optimised sensor system?

?

Most predictive maintenance systems are fairly simple to understand. Data collection on individual machinery is performed by strategically placed smart sensors, which then communicate this data to a central control system for analysis by specialist software. This helps to find potential issues and identify actionable insights before problems occur.

?

Sensor selection ??

Making the most out of a predictive maintenance system begins with the initial collected data. Choosing the right sensor ensures faults are picked up quickly and accurately, with ample notice given for proper repair.

?

A PdM sensor often contains an accelerometer. Able to determine acceleration in one or more axes, these sensors provide tilt and inclination measurements and impact recognition functionalities. Converting acceleration data into vibrational information offers additional insights, as vibration changes are often considered as an early potential fault indicator.

?

By comparing near real-time vibrational information with historical data, sensor systems can identify and flag potential issues. This, combined with the relative low cost and compact size of a MEMS-based acceleration sensor, makes it ideal for PdM systems.

?

But there are times when vibration sensors alone cannot register the fault. In this case, using another sensor type is preferable, either in conjunction with an accelerometer or as a replacement.

?

For example, temperature sensors could help to detect temperature increases caused by excessive loads or start/stop procedures. Where bearings have become worn or damaged, friction is likely to increase, which again can cause a detectable temperature change. Left unchecked, this can cause additional damage to equipment and lead to inefficient processes.

Optimising electronics ??

Once you’ve chosen a suitable sensor type, there’s still more that can be done to deliver an outstanding system, which can be done through optimising the electronics at the very core of the sensor.

?

Many traditional sensor types collect analogue values, which must be converted into a compatible digital format before being communicated to other equipment in the sensor system for analysis. These signal conversion and communication processes may be performed by one or more standard ICs. But what if you require a more optimised predictive maintenance sensor?

?

Manufacturers should instead look to an application specific IC, or ASIC, a chip designed exclusively for an individual customer or application. Within the predictive maintenance context, this could be a chip with sensor-specific processes and conditioning.

?

This means it can offer:

-????????? More accurate signal conversion with reduced noise

-????????? A flexible design process allowing unneeded components to be removed

-????????? More precise sensor data, often with a smaller footprint

Optimising sensor systems from chip level guarantees a predictive maintenance system offering the very latest sensing capabilities, and its users with the benefits of a more efficient, proactive maintenance system.

?

Swindon has over 40 years’ experience in ASIC design. To learn about working with Swindon, book a free consultation.