Published: 14 October, 2014
Kamalina Srikant* looks at improving your maintenance strategy with online condition monitoring.
Everyone who owns a car is familiar with routine maintenance to perform oil changes, tire alignments, and inspection of critical systems such as the brakes. Car manufacturers recommend a scheduled maintenance strategy based on number of months or number of miles driven. But it turns out that you could be changing your oil too frequently, wasting time and money, or if you're driving in dusty conditions, not frequently enough. Scheduled maintenance can even miss other impending issues with your car. Real-time diagnostics can help car owners optimise maintenance for cost and time. While the car industry still has a long way to go, operators of industrial capital equipment cannot depend solely on scheduled maintenance and are moving towards a predictive maintenance strategy.
In today’s fast-evolving global economy, companies that rely on capital rotating assets are facing increased reliability concerns, and unexpected downtime and maintenance can lead to significant cost and safety repercussions to the extent where it can easily affect a company’s bottom line. More than ever, organisations need a dependable predictive maintenance strategy to ensure reliable production and customer satisfaction. Condition-based maintenance solutions help alleviate these risks and can lead to millions of dollars in ROI.
The Electrical Power Research Institute has calculated comparative maintenance costs in US dollars per horsepower (HP) for different maintenance strategies. According to the research, a scheduled maintenance strategy is the most expensive at $24.00 (£15) per HP. A reactive maintenance strategy is the second most costly at $17.00 (£11) per HP, but it can also be dangerous. A predictive maintenance strategy is the most cost-effective at only $9.00 (£6) per HP and cuts the risk of secondary equipment and human damage from catastrophic failures.
1. What Is condition monitoring?
Condition monitoring is an aspect of predictive maintenance that provides all of the information you need to make maintenance scheduling decisions. It involves comparing key measurement indicators, such as vibration and power consumption, to baseline normal behavior to determine if there is any equipment health degradation. It consists of data collection, signal processing, and analysis to provide a complete picture of machine health.
2. Online automated condition monitoring versus manual diagnostics
Traditionally, condition monitoring is applied through routine manual diagnostic rounds. However, trends such as lower cost sensors and automated monitoring systems and the emergence of Big Data analytics are fueling the adoption of automated solutions. Applying online condition monitoring to both critical rotating assets, such as large turbines in power generation, and auxiliary rotating machinery, such as compressors, pumps, and fans, in a given production environment provides the greatest insight into the overall reliability of the fleet of assets or plant, helping a company to thoroughly understand their operations and make business decisions.
For large, expensive capital equipment and rotating machinery, the cost of implementing an online condition monitoring solution is easily justified.
• The most important benefit is an increase in revenue, which comes from maximum uptime and optimal efficiency of production machinery. Properly functioning machines guarantee maximal yield, and by monitoring production machines, you can also detect flaws in product output based on machine behavior, reducing scrap and raw material use while increasing product quality.
• You can also see reduced costs through such a system. With strategic repairs, the operating and maintenance costs of machines with a condition monitoring system can significantly drop. The system can also identify developing faults with enough lead time to properly schedule maintenance during planned downtimes, avoiding expensive plant shutdowns.
• By monitoring various performance parameters, the condition monitoring system can actually help warn of impending risks of failure and help to avoid serious injury. Online monitoring systems also prevent workers from having to enter dangerous environments to take measurements.
There are numerous reasons for companies to move from offline, manual data collection to online, automated monitoring and diagnostics for their predictive maintenance process.
• Workforce optimisation: Manual diagnostic rounds can be extremely time-consuming and require significant travel and setup time, leaving less time for specialists to actually analyse data and assess required maintenance. In addition, many industries are reporting that qualified predictive maintenance and vibration specialists are nearing retirement. Online condition monitoring helps ensure that specialised personnel are spending maximal time on the highest value tasks.
• Fewer gaps in data: When performing manual rounds to collect data, companies typically can only collect a few measurements for any given piece of machinery a month, if at all. A typical power generation utility takes over 60,000 measurements per month. In certain cases, line operators manually noting data values can make mistakes or even copy previous results. Online monitoring removes these errors and helps ensure continuous data collection.
• Improved diagnostics: By using a single database, more historical trend and baseline data is available for predicting faults with greater statistical significance. In addition, with manual diagnostics, the interpretation of a fault is often based on the experience and knowledge of a specialist, and this experience can significantly differ from one specialist to the next.
3. Considerations for Choosing a Condition Monitoring System
Before choosing a condition monitoring system, it is important to understand what types of machines and which failure modes need to be monitored. The breadth and number of machines and the types of measurements needed to detect the failures will form a basis for your decision.
Once this is identified, consider the following when choosing a vendor for your condition monitoring solution:
• The flexibility of the solution to scale with your evolving needs, such as support for new types of algorithms, support for a wide variety of I/O and emerging sensors, and ability to scale to large numbers of systems.
• The openness of the platform to allow you to gain access to the raw engineering measurements and extend the solution to meet your maintenance program requirements.
• Interoperability with third-party hardware and software packages so that you can integrate with existing CMMS and ERP systems and any database historians or process management enterprise software used.
• The breadth and quality of the company’s product offering, including ruggedness of hardware and number of algorithms available.
• The price of the monitoring hardware and the software solution, and how well it allows you to scale your online condition monitoring solution to cover the bulk of your rotating machinery assets.
• The services offered to help facilitate your end-to-end solution from your asset to your IT infrastructure, either directly or through a network of partners.
When implementing a large-scale condition monitoring system, there are three main technology considerations that come into play. The first is data management, which involves using an appropriate data structure, database considerations for easily mining data, alarming capabilities, and implementing an aging strategy.
The second is data analytics, which includes application-specific algorithms and higher-level predictive analytics or prognostics. It involves both real-time decisions and embedded intelligence closer to the sensor source and performing at-rest data analytics on servers using aggregated data from multiple machines.
As the number of data acquisition or monitoring systems increases, data management and data analytics become more complex and a third consideration becomes increasingly important: systems management. Remotely managing large numbers of monitoring systems helps to increase reliability, serviceability, and availability of the overall solution, so that you can perform tasks such as viewing the health of all of your systems, connecting to the network and acquiring accurate data, and remotely configuring channels and uploading firmware application images to your systems.
*Kamalina Srikant is responsible for product management and market development for condition monitoring Solutions at NI.
For further information please visit: www.ni.com