Maintaining machines and equipment at peak performance levels can be a constant battle. Many asset managers are turning to predictive maintenance as a tactic to take the guesswork out of asset upkeep.
By using data and analytics, predictive maintenance helps asset managers determine the best time to replace equipment and send in the engineers, which is usually just before complete failure occurs, but not unnecessarily early.
Shifting to this approach can require a change in mindset and adapting your usual work habits to new technology; once you've decided to take the leap, here are four steps to take for a smooth transition.
Allowing critical assets to run to failure can exacerbate issues, pose a danger to operators, and result in increased downtime; therefore, it is best to avoid this strategy for your most important assets.
But which equipment will benefit from a predictive approach, and which should you replace as needed on a time-based or reactive-maintenance model?
To determine this, you must first find out what information has been and is currently being collected for each asset. For example, do you have monitoring information, such as sensor data, that can be analyzed? Do you have an inspection plan? Have you recorded all maintenance history? From this information, you can determine how critical each asset is and whether you have enough information for adequate analysis.
Data is crucial to foreseeing potential issues, so assets with plentiful, good-quality inspection and sensor data, along with past maintenance information, are ideal for predictive maintenance.
From here, you can put together a plan that allows you to increase the intervals between maintenance events and reduce your overall costs, as well as sustain and improve plant reliability.
Understanding the current condition of your equipment alone is not enough. Only by combining real-time data from sensors—the more the better—with historical trends and other data sets, such as age, inspection records, usage, and output, can you build a full picture of the health of each asset.
Typically, you can use existing sensors for this kind of monitoring. Occasionally, however, the sensors installed in the factory are inadequate to anticipate all the failure modes. In this case, you may need to add additional ones, such as partial discharge or flux probes.
It's important to be able to make sense of monitoring data and look for very specific indicators to recognize patterns. For example, if a temperature on a device increases suddenly, then you know there's a problem. Similarly, if a particular signal stays within a normal range, but is still behaving differently from the way it usually behaves, you should investigate the new pattern.
The value from this pattern-recognition approach lies in taking that information and drawing a conclusion about what's happening in the physical world. Over time, this kind of subtle monitoring allows you to notice sooner when a particular component has an issue.
In 2016 alone, the analysis firm IDC found that companies spent $2.4 trillion on IT products and services. IDC also predicts that this number will increase to $2.7 trillion by 2020.
If you're making big investments in predictive maintenance and sensor analysis, it's important to make sure you get the most out of it.
As an asset manager, you may decide to do analysis and monitoring in-house, which requires building your own maintenance and data center and training your engineers to study and interpret the data. You can also outsource that work to a partner who specializes in it.
A partner does all the analysis for you, using specialized tools and expertise, and then informs you when a problem occurs. Often, selecting the right strategy and partner is the key to a successful predictive-maintenance plan.
Furthermore, when monitoring becomes more important and the technology more sophisticated, outsourcing to a specialist has several advantages. But what should you look for in a partner?
Typically, partners will stay in touch periodically, giving you daily, weekly, or monthly updates, so practical issues such as time zone and language are important. However, finding a company that knows what matters to your operation is paramount.
For the best results, any partner you choose should have a large data set and monitoring fleet, excellent technical expertise, and an expansive portfolio of analytics. They should also be able to help you carefully develop a reliable maintenance strategy, which may involve helping you select which assets to put on your plan.
You'll be working closely with any partner you choose, so make sure the people behind the technology are the right fit for your organization. While vetting partners, take the time to ask questions about the companies' culture, response times, and preferred communication channels.
Managing spare parts and supplies can be a challenge—one that's exacerbated when dealing with many different types of equipment. Spares can also be expensive and a significant segment of your maintenance budget. Therefore, lastly, managing your spare-part inventory is a good idea to improve the success and cost-effectiveness of predictive maintenance.
Transferring to a new maintenance approach can seem daunting initially. But once you find the right partner, together you can plan how to move forward. Remember to discuss what assets to monitor, what data is needed, if any additional monitoring equipment is required, and how to approach inventory management. Besides this, all that's necessary is a fresh mindset and a willingness to take a new approach to improve your plant.
Power plant management challenges can keep you from getting a restful night, unless you know how to tackle them.
Renewable energy's share of the market is growing, but what will it take for zero-carbon power generation to replace fossil fuels as the leading global energy supplier—will it be more than price parity?
MYAs between OEMs and operators are complex. Here's a look at contract management best practices.