From condition monitoring to predictive maintenance
Published by Jonathan Rowland,
Anders Omnell, Director Product Management, Hägglunds Products and Solutions.
Today’s industry is marked by a steady buzz about connectivity and the Internet of Things, frequently summarised as Industry 4.0. New solutions appear all the time, offering a seemingly endless range of technical possibilities. However, it can be difficult to find examples of how these are being implemented.
For many businesses, the greatest potential in Industry 4.0 lies in condition monitoring. Yet even here, customers have generally been slow to act. As with any data-driven service, condition monitoring is only attractive if it offers both security and value. But the real value is created when condition monitoring becomes predictive maintenance.
Condition monitoring is where it begins
In simple condition monitoring, sensors are used to watch over equipment and alert the owner to any changes in operating parameters, such as RPMs, temperature, and pressure. A solution can be created with relatively few sensors and a minimal amount of standalone hardware, at a cost far less than that of a production stop.
The information the sensors provide, however, can be used for much more than alerts. Analysed properly, it can help determine when a machine is at risk of breaking down, so that corrective maintenance can be planned in advance.
Predictive maintenance has many advantages: reduced need for onsite spare parts; fewer demands on service personnel; lower energy consumption; and far more reliable production. So why aren’t we seeing more of it?
Customer concerns about data security
For many customers, the main obstacle on the path to these benefits is the security of the data and the way in which it will be used. Customers worry that sensitive information might be lost or find its way inadvertently into the wrong hands.
These are valid concerns that should be taken seriously. Naturally, the data’s use should be regulated in a non-disclosure agreement, signed between customer and supplier. But the supplier must also take responsibility for having sufficient technical safeguards in place – and be able to explain exactly how those safeguards protect the customer.
Even so, it is important to remember that the situation is generally similar to a manufacturer or workshop accessing a car’s stored driving data. If no data is shared, the possibilities for service – and the potential benefits for the owner – are more limited.
Delivering data or delivering value?
While security issues can be solved by suppliers that take the responsibility, it is more complex to actually bring value to the customer. Sensors are inexpensive and simple to install. But the difficulty lies in turning the gathered data into useful information.
In itself, condition monitoring does little more than supply customers with data, which is often difficult interpret. This is not a solution, but rather an additional headache. For customers to get the reliability and predictability they seek, the data must first be filtered and analysed by experts. Instead of data, customers should receive regular reports and clear insights that help them take the service actions needed.
This is the key behind those solutions that have been successful on the market. Achieving predictive maintenance requires an integration of technology and human expertise, where algorithms supply the basis for smarter decisions.
Creating a machine health index
One of the suppliers that has been successful in this regard is Bosch Rexroth, whose predictive maintenance solution for Hägglunds direct drive systems, Hägglunds CMp, supports a rapidly expanding group of customers worldwide.
Hägglunds CMp establishes a secure link to the customer’s drive system, where sensors are installed throughout. Bosch Rexroth’s internal firewalls protect all information that connects transmitted data with the customer, who has access via a gated cloud interface. While in transit, the data itself is meaningless, since the raw numbers cannot be paired with a sensor’s location.
Meaning is created by a proprietary analysis tool called ODiN, which Hägglunds CMp uses to interpret the sensor readings. During an initial machine learning phase, ODiN creates a health index that provides an accurate picture of the drive system’s normal state. This index is then used by Bosch Rexroth experts to assess and regularly report on the drive’s condition, as well as to spot any major divergences. If the latter appear, the customer receives not just an alarm, but also a concrete recommendation for actions to be taken.
“It’s the combination of health index and expert analysis that makes our solution work for the customer,” said Mattias Ljungdahl, Application Manager, Hägglunds Inside Intelligence. “With the health index, our analysts can not only see the trends, but also understand and predict the drive behaviour for the customer. And because our algorithm is constantly learning, the solution gets better the more it works with an application.
Integration that strengthens service
The benefits of such an integrated solution can be enormous, says Ljungdahl, both for the customer and for the supplier. “Besides helping customers predict and avoid unplanned operating stops, we can use Hägglunds CMp to help them optimise their systems for better efficiency,” he says. “And at the same time, the solution increases the efficiency of our own service organisation.”
As Ljungdahl points out, service becomes faster, simpler, and more effective when the supplier has detailed knowledge of the customer’s drive system. It is not only the alarms that make a difference, but also the regular reports that form a long-term basis for understanding.
“When our service technicians go to a customer with Hägglunds CMp, it’s not just the problem they know before they get there,” Ljungdahl explains. “They’re familiar with the machine and how the application works normally, which makes it easier for them to get things working properly.”
This highlights again the importance of both technical and human components to a successful solution. “The value is not the data, but in the ability to interpret the data and use it to act,” concluded Ljungdahl. “Predictive maintenance is not about sensors or even algorithms. It’s about boosting the customer’s reliability and profitability through insights and expertise.”
Read the article online at: https://www.worldcement.com/product-news/01052018/from-condition-monitoring-to-predictive-maintenance/
You might also like
Selecting the right substitution
Tahir Abbas and Michalis Akritopoulos, Cinar Ltd., discuss the key factors cement plants should consider when increasing the substitution rate of alternative fuels in their production process.