Turing’s Grid: from Smart Grids to Digital Energy Platforms
It may come as a surprise that, for a person in ICT and still amazed by its advances, I believe that the greatest engineering achievement of the 20th Century has been the pervasive deployment of electricity to power our societies. We take for granted that a lamp will light when we flick a switch, and we don’t remotely think about the complex machinery necessary to enable it. Electrification has provided “the blood” for global economic growth for more than a century, perhaps without due recognition.
While electricity has previously created the leap-forward in our economies, today, the utilities sector is being left behind by other sectors in this digital age. Some people joke that if Edison were transported to our present day, he would be able to understand and manage our electricity grid. Obviously, this is an exaggeration, but it does reflect a pinch of truth. Utilities are somehow “laggards” in the digitalization race. It is clear that upgrading the whole grid for a fully supervised and automated network is a complex, lengthy endeavor. But, as the impact of digitalization is showing, it is not the infrastructure what matters as much. It is the information that you gather from it and how you “squeeze” it for value what makes the difference.
Through our work developing the Atos Industrial Analytical Platform for Smart Grid Applications, we’ve found that by capturing the data from all of utilities’ enterprise systems and external data sources, there aredozens of use cases that “reinvent” business. This could be realized using advanced analytics for instance, bringing important economic returns to Utilities all across their value chain.
Let’s look at a concrete example, one of the first implemented in our platform, Revenue Protection. At first it might not seem like the most interesting of analytic use cases, but the economic case around it is solid. It is estimated that, in advanced economies, around 1% of all electricity is lost to fraud and other types of Non-Technical Losses (NTL), with higher numbers in emerging countries and regions like India or Latin America. They may seem like low percentages, but actually that translates to losses of $89.3 billion a year worldwide.
In our Revenue Protection solution, we combine utility data (from smart and analog meters, but also other systems like workforce management) with external data (weather, real estate information) and, with a “toolkit” of analytic methods, we provide better predictions on potential NTL cases. That reduces the number of inspections, while providing better recovery of the unpaid energy. Investment in a Revenue Protection project is recovered in less than a year, a welcome benefit compared to the long term ROI that utilities experience for their infrastructure projects.
However, the key point is not thinking about specific data analytic solutions, but about data analytic platforms. Many use cases operate over similar data sets, and even higher-level analytic components may be reusable across them. What’s more, you reuse knowledge from experts, the hardest resource to find in analytics.
If we look into the evolution of analytic platforms for utilities, the future is truly exciting. There will be more powerful analytic capabilities, like Deep Learning, that will propel Predictive Analytics (“what I can expect”) into true Prescriptive Analytics (“what I have to do”). We foresee the integration of Analytics with High Performance Computing , for scenarios combining prediction and simulation. We’ll also experience the emergence of Distributed Analytics, moving analytics “into the field”, with deeper integration with the network, and a focus in “strict” real-time requirements.
Without doubt, Smart Grids are transformative for utilities. But, as we saw in our introduction, even state-of-the-art infrastructure will become “invisible” to users. It will be the advanced exploitation of the data generated that will become the main driver of competitive advantage for utilities. They will have to evolve their concept of the Smart Grid, currently focused on infrastructure, towards “Digital Energy Platforms”, where energy and information are fully combined. Or, paraphrasing the earlier joke, a Smart Grid where Turing would feel more comfortable than Edison.