Digitalizing predictive maintenance to improve asset management
In 2017, KBC and research firm IQPC teamed up to survey more than 100 operations leaders across the refining and petrochemicals industries. The goal was to gain a deeper understanding of how leaders are responding to current market conditions and setting themselves on a more sustainable path. Although the survey focused on refining and petrochemicals, its findings are applicable to a broad range of
process industries.
KBC’s research uncovered three primary areas of common interest, reflecting a changing approach to operational excellence in general with implications for asset management, and shaped by emerging technologies and an increasingly volatile and competitive landscape.
1. Operational excellence and asset management are important.
Seventy-nine percent of respondents agreed an operational excellence mindset would result in a safer, more reliable and more profitable operation (see Figure 2). However, there are barriers hindering implementation, with only 25 percent of those surveyed feeling their company has a clear vision for achieving operational excellence. Factors identified as barriers include internal organizational barriers, lack of change management capability and lack of understanding of what operational excellence really looks like.
2. A new approach is required to deal with concerns.
Refining and petrochemical companies were mostly concerned with global economy/market volatility, market competition and increased cost of regulatory compliance. New trends such as changing workforce demographics, cybersecurity and technological disruption are also emerging. Across the board, companies reported they were in the "planning, but not yet prepared" stage of readiness for these new trends.
3. Adoption of digital technologies is slow.
Nearly a third of respondents said that "we are quite conservative; our focus will be on the adoption of proven technologies, with new technology mostly on trial." Encouragingly, many respondents reported high levels of maturity in the adoption of technologies such as advanced process control and maintenance/asset integrity systems. But 65 percent of companies said they were not ready to adopt artificial intelligence (AI), with 41 percent and 38 percent saying they weren’t ready to implement solutions built around the Industrial Internet of Things (IIoT) and big data, respectively.
For asset management to be most effective, this mindset must change or industrial plants and facilities risk being left behind. Digital technologies will drive best practices by shifting asset maintenance from preventive — based on a calendar or hours of operation — to predictive, based on analysis of digital data.
Digitalizing predictive maintenance to improve asset management
When it comes to asset management, the aim is to keep the asset up and running safely at maximum efficiency while keeping asset maintenance expenditures in check. There is inherent waste with just preventive maintenance because it inevitably results in maintenance being performed too frequently — driving up costs — or not soon enough — degrading performance in the best case and causing downtime in the worst.
In contrast, predictive maintenance enables asset management through acquisition and analysis of data. This is used to optimize maintenance while minimizing expenditure and downtime. The more effectively this acquisition and analysis of data can be done, the greater the benefit. Tying this back to the survey, the most effective predictive maintenance programs make use of the same technologies respondents said they were not at a stage to adopt: AI, the IIoT and big data.
This disconnect can be addressed by working with suppliers and other resources to implement modern predictive maintenance programs. Before looking at the details of such a program, this article compares the two main types of maintenance performed at process plants and facilities.
Preventive versus predictive maintenance
Preventive maintenance is mainly time-based. It prevents trouble by replacing parts based on factors such as period of use, number of operations and threshold of scheduled maintenance. Predictive maintenance is mainly condition-based. It prevents potential trouble by gathering and analyzing data related to environmental and operating factors, allowing maintenance to be performed optimally.
Demand for predictive maintenance is growing because it reduces downtime and cuts maintenance costs. By maintaining on a condition basis, intervals between maintenance can be increased in some cases, and maintenance can be planned with small problems fixed quickly before they escalate into major issues. This leads to fewer shutdowns and lower costs.
In the past, part and component life was optimized by considering indirect failure factors. This was done by experienced technicians judging the signs of failure based on expertise that varied significantly from one person to the next. However, with technological advancements such as the popularization of small wireless sensors delivering diagnostics, it has become possible to gather and analyze a greater amount of data, which reduces the need for interpretation by experts.
In most plants, preventive and predictive maintenance work together. Preventive maintenance is performed on less critical assets for which the cost to measure, collect and analyze operating data exceeds the anticipated benefits. Predictive maintenance is used for high-value and critical assets for which the benefits of increased uptime and lower maintenance exceed the implementation and operation costs of a predictive maintenance solution.
With advanced predictive maintenance, prediction is guided by measured data. This is enabled by digital innovations such as AI, IIoT and big data, which maximize the reliability and availability of plant assets by achieving higher predictability of operation. Proper utilization of these digital technologies lowers the cost of predictive maintenance implementation for each asset, allowing plants to move more assets from a preventive to a predictive maintenance regime.
Pillars of predictive maintenance
Advanced predictive maintenance systems need an effective infrastructure based on the IIoT to ensure they provide predictable and optimized results. The three main pillars of an IIoT-based predictive maintenance system are:
- Secure remote infrastructure for system monitoring
- Global information sharing system of big data for asset management
- Cloud-based web applications for remote support of plant engineers
From this basis, predictive maintenance can be conducted by using remote support to increase prediction accuracy. This requires active use of an internet-based shared information database and the latest technologies such as IIoT, AI, machine learning, environmental diagnosis and compact wireless sensors.
Specific tools for an effective predictive maintenance program built on the foundation of these three pillars include:
Remote system monitoring service
By gathering maintenance and environmental information, a supplier can check symptoms to spot any trouble factors. The gathered data is analyzed, evaluated and is periodically submitted as a report to the end user. This report can be used for maintenance planning throughout the system’s life cycle. Relying on outside experts in this manner can provide a valuable supplement to plant personnel expertise.
Installation environment monitoring service/environmental diagnosis (online diagnosis unit)
A supplier can gather data related to temperature, humidity, corrosion rate and floating dust. With this data, it can create a hardware health evaluation report based on the data acquired from the installation environmental factors. This allows end users to easily grasp any improvements needed, or the replacement cycle of specific products.
Network health check service
A supplier can collect and analyze network traffic data and detect any unusual traffic in the plant network, along with possible indications of potential cyberattacks. This type of service supports the network responsible for delivering the digital data required to implement a predictive maintenance program.
Plant resource manager (PRM)
A PRM is a key platform because it can improve operations and maintenance by maximizing the reliability and availability of plant assets, thus achieving greater predictability. PRM is a tool that gives end users online access to all their field devices via a field digital network so they can carry out essential management tasks, such as changing device parameters. These types of adjustments, such as reducing pump speed, can stretch asset life to the next scheduled maintenance period.
Small wireless sensors
These types of sensors are a key component of any IIoT implementation because they are the "things" in the IIoT, providing the data required for analysis (see Figure 1). For example, they can provide vibration measurements on rotating equipment, with this data used for predictive maintenance.
Communication support devices
These devices provide support in the field from a remote location (see top photo). Operation should be intuitive and provide visual transmission by video calls and augmented reality, as well as information sharing by sending images and text. This improves the efficiency of maintenance work, reduces losses due to mistakes and facilitates safe and worry-free plant maintenance.
Security operations center
Networks and systems are becoming more complex, and end users are facing many challenges when it comes to ensuring the security of operation and assets. These include the need for sophisticated protection, skilled security resources and real-time monitoring. An operations center offers security monitoring as a managed service, helping end users recover from security incidents as quickly as possible with the collaboration of the affiliate’s service organization with predictive technology.
Conclusion
Digital technologies provide a wider and more cost-effective use of predictive maintenance systems, allowing companies to further their aims of operational excellence.
These tools should be implemented as part of a life cycle execution plan. This plan will rely on a global information-sharing system using big data from various sources including the installed base control and monitoring system, asset management system and maintenance system.
Maintenance policy, asset health condition and other concerns can be addressed using the installed knowledge database to create the life cycle execution plan. The plan should include preventive maintenance, predictive maintenance, obsolescence management and spare parts management.
Executing this plan will result in more uptime, reduced maintenance costs and more efficient operation of critical plant assets.
References
- Operational Excellence in Refining and Petrochemicals: https://www.kbc.global/insights/whitepapers/operational-excellence-in-refining-and-petrochemicals
- Changes in Data Analysis Technology in the Manufacturing Industry in the Midst of the Third Artificial Intelligence Boom: https://web-material3.yokogawa.com/rd-te-r05901-001.pdf
- Machine Learning Applied to Sensor Data Analysis: https://web-material3.yokogawa.com/rd-te-r06001-008.pdf
- Efficient Field Communication with Augmented Reality: https://web-material3.yokogawa.com/rd-te-r06001-005.pdf