How Can We Profit from Operational Monitoring of Key Smart Grid Components?
- Written by Andrea Cavallini, Gian Carlo Montanari and Peter Morshuis
The power system infrastructures in Europe and the United States have attained a venerable age, and as a result, many components are due for replacement. At the same time, the way in which components are operated changes drastically in a smart grid setting. How then can we ensure sufficient reliability for the grid of tomorrow? We claim that operational monitoring of key components is a necessary requirement.
The power system infrastructure is aging and many components are operating beyond their design life. There is a growing risk of unplanned outages occurring more regularly and of failure rates increasing to a point where the industry will be unable to provide replacements fast enough. Therefore, it is crucial that critical components of the network be monitored continuously to prevent unplanned outages, while at the same time their lives are extended without sacrificing a high level of reliability.
The way we operate the grid is changing. At the distribution level, smart grids are characterized by bidirectional and intermittent power flows. Large-scale introduction of power electronics has led to voltage wave shapes that differ from the 50/60 Hz sinusoidal form for which the grid components were originally designed.
New eco-friendly materials, such as natural ester oils for transformers, SF6/N2 mixtures for GIS, and a variety of biodegradable solid polymers, are being introduced into grid components. How will these "green" components perform in the long term?
What are the effects of these changes? Do they affect reliability, and if yes, how?
Over the past twenty years, extensive work has been done to develop mainly off-line diagnostic tools to assess the condition of power grid components. This work has yielded diagnostic markers, which are used to trigger maintenance actions and to evaluate failure risk. Now, the time is ripe to move on towards systems that can be used to monitor components in operational settings. The main drivers are minimizing unplanned outages and optimizing the use of the grid components.
A host of new technologies have come onto the diagnostic systems market lately. They are dramatically changing our approach to the diagnosis of grid components. Let us recall two examples. For the assessment of the quality of the electrical insulation of high voltage cable systems, new partial discharge detection (PD) technologies enable us to measure PD online, in noisy environments. This alerts operators to electrical insulation degradation and increases the reliability of systems.
Another breakthrough has been the introduction of systems that are able to carry out dissolved gas analysis (DGA) on operating power transformers. Again, the reliability is increased in comparison to the situation in which DGA is performed only once a year when the transformer is taken off-line.
Such new systems are often endowed with advanced communication tools, which enable a bidirectional flow of information with SCADA centers. Thus, they permit the condition of the apparatus to be monitored continuously, and the apparatus to be reset to interact differently with the surrounding environment. Alarms, for example, can be reprogrammed.
Eventually, a large number of data processing techniques, from artificial intelligence to data mining, coupled with the knowledge acquired from diagnostic markers, will enable operators to treat the data flow coming from an apparatus in a smart way. The data flow can be reduced to very simple information, such as a traffic light, so that in a SCADA center, attention can be focused on a piece of vital equipment only in critical conditions.
Such data processing techniques also open the way to sophisticated correlation analysis of data streams from diagnostic systems monitoring different quantities. This enables us to exploit synergies in the complementary information brought by these systems.
For a transformer, PD and dissolved gas analysis, along with thermal profiles, can be analyzed jointly to produce a more accurate diagnosis.
Systems that are able to operate in this way could be properly termed Smart Grid Global Monitoring Systems (SGGMS). SGGMS are the proper answer to the questions above: They can optimize revenues for the stakeholders, while keeping acceptable reliability levels. Yet, though SGGMS systems are already available on the market, they are still not commonly employed. The reasons can be twofold.
On one hand, the application of these technologies has been limited so far because there are not a lot of business case studies demonstrating their benefits. It is probably time to take a leap forward, appreciating that the benefits—such as avoiding outages and exploiting grid component—can be very rewarding compared to the costs.
On the other hand, the business-oriented grid management associated with competition in the power sector has produced a system approach to the grid where a transformer, for instance, is somehow considered as a functional block. In a world where component technologies and the operational conditions they are subjected to are changing in dramatic ways, it is probably time to invest more in high-tech condition monitoring systems to replace time-based approaches and obviate resort to "experts"—a less and less available item! (Diagnostic measurements tend to be performed at regular times, with the apparatus off-service. With monitoring, these measurements can be performed continuously.)
How do we begin the process of introducing on-line monitoring? It makes sense to start with the "low hanging fruit," to make use of sensors and technology already installed or available. If we focus on a modular approach, we can gradually build a smart grid global monitoring system, while convincing stakeholders that the approach works and actually provides the promised range of benefits.