Portfolio Management in Energy Trading

Central Energy Trade Hungary Group works in energy trading, innovative investments and renewable energy development. Based on years of trading experience, they see the future in supporting their processes with modern software and automating manual steps.

Their portfolio covers consumers, producers and wholesale partners, where production and consumption are represented by units, so called PODs or points of delivery. They asked Dmlab to build a complex data collection, portfolio management and trading solution on top of this structure. We ran the work in the phases we use on every data project, from understanding the business problem to putting the model into operation.

The software has to manage data for every client and point of delivery, collect production and consumption readings, weather and price data in real time, and forecast the output of solar farms and the consumption of individual clients.

On top of that it has to trade automatically on the energy exchange based on those forecasts, generate invoices for consumer clients, support strategic planning through dashboards, and finally monitor its own operation and raise an alert when something goes wrong.

The solution draws on many data sources. Solar farms send production data by e-mail, while real time consumption readings are collected directly from the relevant servers. Dedicated collector modules bring in daily energy prices, weather forecasts and calendar information on schedule, and the trading module writes the results of its own transactions back into the database.

Production and consumption trends are shaped by weather, public holidays and historical patterns alike. The software therefore builds predictive models that use machine learning to forecast expected volumes, down to individual production and consumption units. These forecasts drive the portfolio balance and the trading decisions.

Using the latest predictions, the system calculates the current and future balance across the entire client base, that is, how much energy has to be bought or sold and when, and from that the monetary value of the portfolio. The trading module settles emerging gaps in intraday trading and automatically files the required balance data.

Thanks to the invoicing integration, invoices for consumer clients are generated according to the agreed pricing, and producer clients automatically receive the data they need to issue their own invoices. Automated e-mails also deliver unit level forecasts to the clients.

The software provides a responsive browser interface where the current trading position and the daily tasks are visible from any device. The Power BI-based dashboards bring together production and consumption patterns, energy prices, weather forecasts, current and future balance predictions, and the data needed for strategic planning, such as available cross border transmission capacity.