Load forecasting, designed to ensure there is ample supply of electricity to meet demand, has long been a focal point for utilities the world over – and it is especially so in the ongoing energy crisis. In South Africa, facing continual load shedding, efforts have intensified to relook at the country’s energy security to meet current and future demand.
Satyajit Dwivedi, Regional Director, EMEAP, Energy Utilities, Mining & Metals, Public Sector, SAS.
Satyajit Dwivedi, Regional Director, EMEAP at SAS, a leader in analytics, notes that the ability to predict customer demand accurately and better manage infrastructure to ensure industry, business, education, healthcare, and households have the electricity they need, is a careful balancing act. On the one hand, a utility must fulfil its service mandate, and on the other, it has a responsibility to keep the infrastructure operational, given capacity limitations in the short, medium, and long term. Dwivedi suggests that all stakeholders in the energy ecosystem can benefit by adopting the latest technologies, including data analysis, to identify how to maximise resources.
In this context, advanced artificial intelligence (AI)-powered energy forecasting tools can help power utilities manage the changing complexities of power markets. More frequent extreme weather events, coupled with the growth of alternative energy resources, are having a significant impact on how utilities and power providers prepare energy forecasts. Adding to this complexity are ongoing changes in market regulations as sustainability imperatives become part of government mandates around the world.
A changing paradigm
Dwivedi says AI-powered energy forecasts can provide scalability and access to actionable insights for stakeholders in power and utility companies who need timely and accurate information. “AI-driven models that leverage neural networks, machine learning, and deep learning are proving to be highly accurate, providing utilities with data and analysis to make better-informed decisions.”
These technologies can automatically integrate any number of variables such as ambient temperature, humidity, heat index for short- and medium-term predictions. When forecasting electricity demand, several variables can change by the day, hour, or minute. AI-driven models can address and capture these conditions to predict sudden changes in the underlying variables accurately, resulting in more accurate forecasts. They can also use macro-economic factors, renewable energy impact, and climate change to estimate long-term demand.
“Data analysis through AI-driven solutions that integrate energy forecasting makes predicting overload conditions, capacity utilisation and network strengthening needs clearer, to build a resilient energy grid,” Dwivedi adds. “In this way, investments can be optimised, taking account of limited budgets. Additionally, leveraging AI and the Internet of Things, effective predictive maintenance strategies can be designed for power plants to increase availability.”
Considering factors such as the economic trajectory, climate change, changing portfolio of end use and the impact of demand side management, the value of accurate long-term load forecasting cannot be overestimated, Dwivedi says. Forecasts with improved accuracy may defer building additional generating units or the need for a long-term power contract. Accurate load forecasts also have budgetary implications for the power supplier. Lead times to build the required generation can involve years to acquire the necessary approvals. The capital cost of building new generating units is incurred by the utility and in turn passed on to customers. Alternatively, if there is not enough generation to meet future demand, the power provider will need to secure long-term power supply contracts separately.
Leveraging business data and analytics for intelligent decision making can therefore assist and support governments and utilities at macro and microeconomic levels. It can empower them to improve planning, investments, maintenance, and service delivery.
Adopting a data- and analytics-driven approach to energy forecasting and demand-side management, also helps power suppliers stay on top of short-term maintenance and take care to cater for future demand.
AI-driven models can provide benefits in areas such as operations, trading, and integrated resource planning. Accurate forecasting models can help power and utility companies produce a significant return on investment and deliver on their operational mandates.
For more information visit: www.sas.com/en_za
