In the context of this article, the term “smart device” is defined one that is programmed to be capable of some independent action. We also use terms such as big data, cloud, neural networks, artificial or augmented intelligence (AI), virtual reality, machine learning (ML), digital transformation, cyber-whatever, and any other number of technical or marketing buzz words for movement or interpretation of information and the results of processing it. What this all comes down to are devices and programs developed by humans to manage large amounts of data that can provide decision-making information or speed pre-programmed decisions that appear to be independent of human intervention. The purpose, from a reliability standpoint, is to take in sensor and other inputs to fuel rapid decisions that are, it is hoped, repeatable. This can benefit all aspects of our lives and livelihoods. It brings new levels of convenience, makes the world smaller, and supports faster decisions across the board.
Looking back at the transition of the early Internet from a global connection for government, university, and independent researchers, and hobbyists to today’s crucial World Wide Web, the original idea was to allow communications through a broad server network. In the beginning, whether via phone-line modem, network modems, or other communication means with addresses into servers, accessing and processing information was only possible if a user had a good connection. As systems and memory sped up and expanded, however, the Internet evolved to the point that ever-greater amounts of data were becoming available, i.e., seemingly too much data. This required new methods and systems for processing that data.
Within the reliability and maintenance community, opportunities that go along with avalanches of data and new collection and processing capabilities range from early detection of defects through inputs from a vast array of sensors to the optimal use of a transitioning workforce. This extends beyond our own facilities as the web spins outward to incorporate energy transmission, including smart-grid inputs, to real-time production information and forecasting to know the expectations of production line loading and logistics. For reliability engineers and asset managers, future capabilities could range from obtaining near-instantaneous condition of equipment and projecting Time to Failure Estimation (TTFE) and Remaining Useful Life (RUL), to predicting the impact on the health of a business and profitability through the knowledge of system loading and distribution.
DECISIONS, DECISIONS, DECISIONS
In past articles of this “smart” series (see links below), we discussed smart-grid and smart-cities programs that are in development. As we look to the future (including the future of U.S. infrastructure), the role of the reliability engineer and asset managers will continue to transition from the aspect of Physical Asset Management. Since the early 2000s, the pendulum has been swinging from the concept of human interaction (primary) to “Industry 4.0,” which involved the removal of human interaction, to the new concept of “Industry 5.0,” which involves the interaction of cyber-systems with human decision-making. Why is this important?
In the world of AI and machine learning, data is fed into either a supervised or unsupervised system, pre-processed, passed through an algorithm (or series of algorithms), and some action or data is produced. This is all designed and programmed by humans who use programming and software systems designed by humans. The data used to train these systems, the determination of how it is pre-processed, and the development and selection of algorithms is done by humans. So is the selection of features in a predictive-maintenance (PdM) system, such as vibration analysis.
In the end, we are working with complex expert systems that process data and produce results, be they good or bad. For example, in 2003, the catastrophic Northeast blackout was caused by a combination of improper tree maintenance, hot days, heavy line loads, and a software defect that, in a few minutes, made decisions to NOT inform the operator about unfolding conditions. In another case, a sensor failed on a critical machine bearing (causing the monitoring system to register a bearing temperature exceeding 25,000 F). The system didn’t take into account that, much like components, sensors also can fail. Human interaction, though, prevented the system from going offline until after a technician confirmed the findings. The overall message here is simple: Effectiveness and accuracy of AI and machine learning is determined by human beings and all that implies.
Going forward, as we continue our global journey to digital transformation, Physical Asset Management (PAM) systems will be receiving information from smart energy sensors, production sensors, condition monitoring, productivity forecasting, scheduling, materials availability, logistics, personnel, and much more that we may not currently envision. While some decision-making may be automated, some information will require interaction. With this, the ability to have the system either assign proactive or corrective actions and materials or inform planners will be implemented.
When we add in self-healing technologies for critical systems, the urgency of corrective actions will lessen, and specialists can become involved based on availability. For unusual tasks, personnel will have access to detailed procedures, information, and checklists on personal devices that will normalize repair and general maintenance processes. Skills and experience will still be required, but the work will be procedurally consistent. Specific information will also be broadcast out to suppliers, vendors, and energy producers, which, in turn, will assist in their planning and scheduling. The result will be less waste, lower costs, and a greater capability of on-demand supply that impacts profitability.
Note that The PAM professional is a critical component of the system and related decision-making at all levels, including those involved in a smart grid and smart cities. When problems do occur, such as the COVID-19 pandemic, resilience does not relate to just one company, but, rather, a broad base of organizations that rely upon instantaneous information to make decisions.
The decision-making goes beyond the operation of an organization, but also the placement of an organization. One of the principles behind industrial engineering is the understanding of resource placement, which includes transportation, warehousing, factories, offices, outlets, and many other aspects of an industrial organization. With the concepts behind smart grid, smart cities, Internet access, highway and other transportation access, decision-making can be far more accurate than it has been.
In the future, new operations will be drawn to locations that can provide specific features, including, among other things: solid interaction between the energy supplier and the customer’s microgrid; traffic and conditions information for planning and routing; and smart-city programs, such as smart-sensoring of road conditions to reduce wear and tear and related costs on company, customer, and employee vehicles. New business operations will also be attracted by services through smart cities, such as the public well-being and neighborhood safety, which can help improve employee productivity and attendance. While many current smart-city initiatives are demonstration projects, the results have been significant regarding communities being able to attract new businesses (which impacts the tax-base), as well as support of existing businesses through infrastructure improvements and shared information. Among other things, that includes providing critical information on existing and planned changes to traffic flow, weather patterns, and resources.
CHALLENGES AND OPPORTUNITIES
Challenges for the future include making the workforce more technology-savvy than it is today, from heeding matters of cybersecurity to understanding the capabilities and limitations of smart technologies. (Note that this reference to “workforce” includes reliability engineers and other PAM Professionals).
As the web of information continues to spin internally within operations and outward to local and surrounding communities and energy suppliers, the ability to forecast equipment and production-line availability increases exponentially. As AI and machine-learning systems become more accepted and sensors become more connected, the ability to project the useful life of an asset becomes academic. At the same time, the data produced by these systems provides the reliability engineer the ability to look at opportunities ranging from asset re-engineering, production improvement recommendations, and new equipment specification improvements to expand the life cycles of existing and future plant assets.TRR
Click On The Following Links For Recent Articles On Smart Technologies/Concepts And Electrical Reliability
“What’s So ‘Smart’ About A Smart Grid? (July 3, 2021)
“Electrical Resiliency & The World Of Microgrids” (July 11, 2021)
“What Makes ‘Smart Cities’ So Smart?” (July 18, 2021)
ABOUT THE AUTHOR
Howard Penrose, Ph.D., CMRP, is Founder and President of Motor Doc LLC, Lombard, IL and, among other things, a Past Chair of the Society for Reliability and Maintenance Professionals, Atlanta (smrp.org). Email him at [email protected], or [email protected], and/or visit motordoc.com.
Tags: reliability, availability, maintenance, RAM, smart cities, smart grid, NIST, microgrids, electrical resiliency, electrical systems, electrical equipment, power plants, power generation, wind turbines, wind energy, solar power, nuclear power, geothermal energy, energy storage systems, vehicle-charging stations, portable generators, cybersecurity
