iiot Smart maintenance and the Industrial Internet of Things

Related Vendors

CeramTec GmbH

EA Elektro-Automatik GmbH & Co. KG

ROHM Semiconductor GmbH

onsemi

The concept of monitoring factory machines for maintenance purposes isn’t new. However, Industrial Internet of Things tools – smart sensors, Big Data, and machine learning – can bring greater insight, more productivity and reduced downtime to predictive maintenance strategies.

For owners and operators of manufacturing businesses, maintenance isn’t just a dull but essential engineering function – it’s a factor that can significantly impact the profitability or otherwise of the entire enterprise. An unexpected failure can immediately inflict expensive damage on the machine concerned and possibly its environment, as well as lost production time. However, this isn’t all; the event could well lead to failed deliveries, disappointed customers, lost sales, and reputational harm.

Maintenance strategy choices

Accordingly, a corrective, or ‘fail and fix’ maintenance strategy - or non-strategy – is no longer an option for most organizations. Condition-based strategies where technicians react to signs of imminent failure are also risky, especially if failure modes can arise without advance warning. A preventive approach, where plant items are regularly replaced or serviced, is better as it can greatly reduce failure rates. Yet it’s not ideal, because technicians may be tempted into premature machine servicing and parts replacement to forestall failures. This unnecessarily increases downtime and production loss, while also inflating part replacement costs.

A better approach is predictive maintenance, where sensors attached to machines are monitored for any deviation beyond predefined limits, so that technicians can be alerted and respond appropriately. Such monitoring can be used for optimizing machine performance during normal operation, as well providing warnings of imminent failure. Accordingly, a well-planned and implemented maintenance strategy can maximize equipment productivity in terms of both performance and uptime.

The idea of machine monitoring is not new. Motors and bearings have long been fitted with vibration and temperature sensors, for example. However, traditional approaches have limitations that could be improved. For example, an onsite control and monitoring system may warn you that a motor is beginning to vibrate or run hot; but it can’t tell whether it should be serviced immediately, or if it could actually run safely and reliably for another couple of months before action becomes essential. And this ‘remaining life’ figure, isn’t simply unknown; it isn’t fixed, either. Its value could depend on many factors, including the exact model of motor, how much it is used, how heavily it is loaded, and how often it is started and stopped. Other less-obvious factors, such as the environment in the production area, may also play a part.

Solar Power

NEXTracker's new control system should keep solar power plants safe in all weather conditions

How the Industrial Internet of Things creates opportunities

This is where the industrial Internet of Things (IIoT), with its Big Data and artificial intelligence (AI) aspects, could contribute to a more sophisticated, effective, and powerful predictive maintenance strategy. Large numbers of affordable sensors could be connected to motors, bearings, valves, and other equipment items around a site. Data from these can be marshaled into a local, or ‘edge’ data center. Computing power within this center could be used to provide immediate responses to urgent situations such as detection of a gas or liquid leak. Other data relating to longer-term trends, such as the temperature and vibration measurements mentioned above, can be sent back to a central or cloud-based facility with more powerful computing capabilities.

Here, motor vibration data over time could be collected from large numbers of machines or processes across the facility, or possibly multiple facilities. Related data such as temperature inputs, manufacturer’s data or service information could also be input. As the historical data base builds up, machine learning software could spot correlations, learn, and infer valuable advice for the facilities’ operators. Ideally, it will advise on how long a motor could be expected to run before maintenance becomes essential, and how this expected run time depends on how the motor is operated. Perhaps, for example, the software could spot that if a type of motor is started and stopped more frequently than usual, then its service life is severely shortened.

Calibration

An introduction to electronic calibration

Implementation example

Some of these concepts are implemented within ABB’s Ability Smart sensor system for motors. This is based on sensors which can be attached or retrofitted to motors without wires, and measure vibration and temperature at regular intervals. Information is encrypted and collected from these sensors using Bluetooth. It is made available to an ABB portal, analyzed, and sent to nominated users’ smartphones running the ABB Smart Sensor app.

This app includes a ‘traffic light’ display of each motor’s current status. Red means there is a critical issue, with failure likely soon; urgent action is needed. Yellow means that operation can continue but the motor should be watched closely and serviced at the next available opportunity. Green means that the motor is fine.

Conclusion

Optimized predictive maintenance strategies offer many benefits. Productivity is improved, as machines achieve significantly improved uptime. Technicians can also be given insight into maintaining and repairing equipment more efficiently, while maintenance parts and labor costs are reduced. Well-scheduled maintenance can also extend equipment lifetime. Improved machine performance can also help cut energy consumption.

Additionally, smart maintenance solutions can include reporting features and procedures that facilitate compliance with maintenance-related standards such as ISO55000.

(ID:46587660)