Now is the ideal time for Eskom to use advanced analytics to improve decision-making and better handle South Africa’s power crisis, says Kroshlen Moodley, GM: Public Sector and Utilities at SAS.
Eskom has always used a degree of analytics to manage power supply. However, the big data it gathers from power plants, transformers, generators and other machinery only reflects events that have already happened, making decisions based on this historical data ineffective.
Looking back vs looking ahead
The biggest problem is that decision-making is reactive rather than proactive.
Take infrastructure maintenance as an example. Many of South Africa’s power plants are operating well beyond their 30- to 35-year lifespans. An increase in demand puts pressure on these plants, which often break down and force Eskom to perform reactive unplanned maintenance. As these events are outside of the normal maintenance cycle, Eskom is forced to delay planned maintenance and to sweat its power assets even more. Anyone in IT will be familiar with the risks of sweating assets, which include a higher probability of breakage, increased costs and sluggish performance.
As a result, Eskom has had to rely on diesel generators to make up for the shortfall when power plants are offline for maintenance. This is also risky as the generators were not meant to supplement daily capacity but rather maintain reserve capacity. They were purchased to supply reserve capacity to cater for fluctuations in demand and to ensure the grid does not slip into a complete failure state.
If this, then that
Advanced analytics can have a massive impact. The big data gathered from Eskom’s machinery, as well as data around area usage and other variables, can be plugged into a statistical model, allowing Eskom to predict future scenarios based on a set of events that have already occurred. This will enable proactive decision-making that will regulate supply and help Eskom to better plan its maintenance schedule.
Relying on reactive data to plan for the future is problematic as decisions on new plants and additional energy resources are only made once supply can no longer meet demand, as is the case in South Africa currently.
Using advanced analytics to forecast demand and supply – and by analysing information from smart meters to determine how much energy a particular areas needs and to forecast how much demand is likely to grow – Eskom will be able to better manage generating capacity to meet demand. When there is a gap in demand, analytics can help Eskom to make better decisions in terms of introducing other electricity sources into the energy mix to close the gap. By analysing weather, seasonal and geographic information, as well as the variables for each alternative energy source, Eskom can make informed decisions on how to create the optimal generation and supply scenario – such as including wind energy from coastal regions and solar energy from inland areas in the mix.
Low-impact load shedding
With advanced analytics, Eskom can also responsibly manage and communicate planned outages. By considering variables such as critical operation times for certain industries and traffic flow, and combining this with the insights gathered from smart meter monitoring, Eskom can figure out how best to supply energy at different times of the day in a way that minimises impact on businesses and citizens. Sentiment will increase as a result, which would calm foreign investor nerves and improve confidence in the economy.
Soaring diesel costs, unplanned maintenance and the knock-on effects of load shedding clearly show that keeping the lights on came at a price. Eskom has to change the game in terms of how it forecasts and plans for demand, and advanced statistical models have a huge role to play.
The good news is that we no longer needed qualified statisticians to build these models. Today’s analytics software is user-friendly and intuitive and can build models for us. It uses a degree of artificial intelligence to meet the requirements of modern-day businesses and presents a number of different scenarios in terms of what might happen. This allows businesses and utilities to make better-informed decisions, saving time and money and optimising processes and resource allocation.