|Kamalina Srikant||Theresa Woodiel|
Kamalina Srikant is responsible for product management and market development for Condition Monitoring Solutions at National Instruments (NI). Ms. Srikant’s work experience extends to the energy, Big Data, consumer goods, and machine vision markets. She holds a bachelor’s degree in mechanical engineering from McGill University in Montreal, Canada.
Theresa Woodiel plays a lead role in advancing the reputation of National Instruments (NI) embedded systems platform and trends within the industry. She holds a master’s degree in technology commercialization from The University of Texas McCombs School of Business.
Q: The phrases condition monitoring and predictive maintenance are commonly used together, but they’re not one and the same. How are condition monitoring and predictive maintenance different?
A: Condition monitoring is a long-established process for monitoring critical assets through measureable and quantifiable parameters such as vibration, temperature and a number of other condition performance indicators. There are multiple approaches to condition monitoring, including predictive maintenance, preventive maintenance, and run-to-failure maintenance. Each approach includes various benefits and risks as outlined below.
Predictive maintenance programs schedule work only if and when mechanical or operational conditions are warranted by (1) monitoring the pumps for excessive vibration, temperature and lubrication concerns or (2) if any other problems are observed over time. When conditions reach a predetermined, unacceptable level the equipment is shut down for repair or for the replacement of damaged parts. This approach prevents a more costly failure occurrence and is beneficial if personnel have the adequate knowledge, skills and time to perform the predictive maintenance work. Essentially, predictive maintenance is designed for scheduling of equipment repair when it’s required and with the least impact on previously scheduled activities.
Preventive maintenance consists of scheduling maintenance activities at predetermined intervals, where operators repair or replace damaged equipment before obvious problems occur. Conducted correctly, operators can prevent maintenance issues before catastrophic problems arise. This approach is beneficial because it works well with equipment that does not run continuously and has properly trained personnel with adequate time to perform the work. The disadvantages include scheduled maintenance taking place too early or too late, resulting in operating errors. In addition, it’s possible reduced production could occur due to potentially unnecessary maintenance. In many cases, there is a possibility of diminished performance through incorrect repair methods, for instance, greasing bearings every month and calling it a preventive maintenance program.
Run-to-failure maintenance makes it possible for operators to run the pump to failure and only repair or replace damaged equipment when obvious problems occur. This approach is most effective if equipment shutdowns do not affect production and if labor and material costs are insignificant. Of course, the disadvantages are that maintenance departments perpetually operate in unplanned crisis management, which results in unexpected production interruptions and high inventory requirements of spare parts for quick repairs. Typically, futile attempts are made to reduce costs by purchasing inexpensive parts, further aggravating the problem. Frequently, maintenance personnel are confronted with solving an onslaught of problems that are never corrected due to a lack of preventive maintenance performance.
Q: Predictive maintenance has been a topic of discussion for more than a handful of years now. Where do we stand in terms of marketplace adoption of predictive maintenance systems? Are these systems commonplace in industrial applications today?
A: The transition of condition monitoring programs has gone through several stages over the last several years. Previously, plant operators manually collected readings on paper that may or may not have been reviewed. Today, condition monitoring is more widely adopted as plant operators collect data of critical assets on handheld electronic devices. Collecting data is the first step in a condition monitoring plan. Then, these devices are returned to a central location that aggregates the data and analyzes it for trends. While this is an improvement to paper readings, operators are still tasked with walking around plants that can house hundreds of critical and non-critical assets, collecting as many as 60,000 readings per month.
Less common today is the integration of a systematic predictive maintenance program that combines condition monitoring across critical and non-critical assets, connecting these distributed assets to ensure they’re networked into the IT architecture. Currently, this type of predictive maintenance program, which seeks 100 percent condition monitoring coverage and visibility into its equipment, is not widely adopted. However, as the pressure to drive down cost increases, power plants will continue to expand condition monitoring across all equipment. This will make it possible for operators to apply analytics to vast amounts of data, so they can provide business insights to their organizations to help improve operational performance. As examples, these improvements could include:
Realizing the “Internet of Things” by taking the condition monitoring world and bridging it with the IT infrastructure.
- Optimizing equipment when it’s needed as opposed to when it’s scheduled and may not be warranted.
Power generation is an industry that has realized the pitfalls of the current, manual approach, and the need to move toward online fleetwide monitoring. Reliability demands and risk of downtime is driving power generation stakeholders to address this problem.
Other industries are at different stages in their realization of the need to address the issue of manual diagnostics. Fortunately, as sensor and computing costs decrease, this enables broader condition monitoring of critical and non-critical (auxiliary) assets.
Q: From an application perspective, what do you see as the keys to success for condition monitoring and predictive maintenance systems? How can end-users ensure their systems live up to their potential?
A: Operations and Maintenance Management personnel play a critical role in binding the distinct parts of condition monitoring and predictive maintenance systems into a cohesive program. These departments are responsible for O&M project implementations and developing the structure while balancing their O&M resources, in addition to serving as the linkage between other departments such as engineering, IT, administration and training.
Given this growing complexity, O&M Management needs to look beyond reliability. This means also considering controlling costs, safety, evaluating and implementing new technologies, as well as opportunities to expand programs across a variety of applications. Historically, the O&M field has focused on reliability as the single most important metric. While reliability is critical, O&M Management personnel also need a real-time, systematic view into the distributed applications throughout a plant that’s remotely accessible to ensure success.
To effectively diagnose all possible faults and to reduce false positive detections, it’s important to incorporate a number of different measurement types into a condition monitoring system—meaning organizations need to seek agnostic hardware and software platforms that are independent of equipment providers to enable this sensor fusion. Often times, control systems are tied to equipment and industrial processes can have multiple types of control systems within their operations. As a consequence, industrial applications are pieced together, similar to a puzzle, when they were never designed to “fit together.”
Q: What are some examples of how condition monitoring is being used for non-maintenance applications in industrial process environments?
A: Condition monitoring-like techniques can be applied in factory acceptance tests of rotating equipment. This is because similar types of measurements that occur in condition monitoring environments are also applicable with factory tests.
Q: What is the coolest/most innovative application of predictive maintenance and/or condition monitoring that you have seen employed in a live application scenario?
A: We worked with PEMEX (PEP) to create a real-time monitoring system for a crude oil distribution system known as Sistema de Monitoreo de Variables Operativas (SIMVO). This company oversees the exploration, production, transportation, and commercialization of oil extracted in Mexico. PEP’s Southern Region Transportation and Distribution Management is responsible for transporting and distributing Olmeca, Istmo and Maya crude oil. In addition, PEP transports and distributes approximately 1.52 million barrels of this oil daily, which represents 43 percent of national production. This volume is equivalent to $3 billion in crude oil. To determine the precise oil volume PEP transports and distributes, it relies on electronic measurement systems installed in the field.
Previously, coordination between the different management teams and separate measurement systems was done manually by phone or email. To enhance coordination between these teams and take advantage of existing measurement systems for the transportation and distribution of crude oil, they realized a need for an integrated and low-cost monitoring system. This application example highlights how industrial organizations need platforms capable of linking to different communication networks through industrial protocols and standards. As an example, OPC is an industrial network that can help protect systems from virus attacks, version incompatibility, and unauthorized access through an off-the-shelf approach that speeds up application development while also plugging into legacy field equipment.
Q: Looking ahead, what do you see as the logical next steps for condition monitoring and predictive maintenance? How will these systems be used in the future in ways that they are not being used today?
A: Figure 1 illustrates the logical next steps for condition monitoring and predictive maintenance. Today the convergence of sensors and cloud computing are making condition monitoring more feasible and accessible throughout an organization. In the future, we anticipate organizations will want to apply prognostics to fix or replace equipment before it breaks. Prognostics focus on predicting the time at which a system or a component will no longer perform its intended function with certainty.
In addition, we expect that power generation facilities, as well as transportation and industrial organizations will require advanced platforms capable of communicating status updates along with alarms that are synchronized over an organizations’ network. Ultimately, when we consider condition monitoring and predictive maintenance, we often think in industrial terms where each piece of equipment will have an IP address and be connected to the Internet of Things. Moving forward, this idea will expand to include smart cities, smart buildings and the smart factories of the future.
|Figure 1. This schematic shows a modern condition monitoring framework with predictive maintenance from the perspective of National Instruments.|
Interview conducted by Matt Migliore, Flow Control’s former director of content. Matt can be reached at Matt@GrandView-Media.com or 610-828-1711.