Swiss energy services company uses machine learning to see the future

If you want to look into the future, sometimes you have to be able to predict it.

Swiss energy services company IWB has a vision of a world with a fully renewable, climate-friendly energy supply. 

Not long ago, though, that goal seemed difficult to conceptualize.

For many years, IWB’s distribution grid supplied customers with electricity exclusively produced by large, centralized power plants. Grid feed-in characteristics were heavily dependent on solar irradiation and could fluctuate. But because electricity consumption was easy to gauge, there was no urgency for measuring current and low voltage power flows.

That changed in 2017 when Swiss voters approved an energy act that would reduce the country’s dependency on fossil fuels by 2050. To help realize this vision, private households began using electric cars and installing heat pumps and photovoltaic (PV) systems, converting light into electricity.

In IWB’s supply area – the Canton, or state, of Basel-Stadt – there are more than 1,700 solar PV systems, with an annual growth of approximately 300.

The new installations shifted the consumption trend, resulting in a higher network load, impacting electric utility company distribution grids.

It was imperative for IWB to assess the level of solar power production. But the measuring solution was complex and required frequent manual adaptions as solar PV systems increased.

Without real-time power measurements, estimated power values were being used. Unfortunately, these were frequently inaccurate.

“The increasing amount of decentralized solar photovoltaic systems represents a challenge for planning and operating our distribution grid,” explained Daniel Grossenbacher, IWB’s Asset Management supervisor. “Without knowing how much power is actually being fed into the grid, a grid-friendly control of this renewable energy source is difficult to achieve.”

If IWB hoped to fulfill its mission of helping to usher in a fully renewable future, an intelligent solution had to be developed to accurately predict solar power production.

Merging the data

IWB Industrielle Werke Basel (IWB) emerged from the private gas industry in 1852, becoming an electric company in 1899. Since 2010, IWB has been an independent company owned by the Canton of Basel-Stadt, supplying the region with electricity, heat, drinking water, and telecom and mobility solutions, as well as producing and selling renewable and CO2-neutral energy. 

As the company conceptualized the best measuring solution, planners understood the importance of integrating existing data from diverse sources.

While scoping and modeling the project, IWB relied on support from SAP’s Global Center of Excellence and Customer Advisory, providing both business and application expertise to organizations engaged in SAP implementations and optimizing existing ones.

Throughout the development process, IWB’s IT team worked closely with the multi-national, enterprise resource planning (ERP) software leader.

“SAP helped us to connect and combine our internal utility-related knowledge… to build a valuable tool which supports us securing the performance of the grid,” said Marcel Holzer, manager of SAP systems at IWB.

Already, SAP’s energy management data (EDM) solution was being used to store information gathered by the smart meters measuring the power production of each of the various solar PV systems.  The problem was that the smart meters were only feeding their data once a day. As a result, total load calculations were not being provided to the grid control center in time.

The new platform would alleviate this dilemma by using machine learning (ML) algorithms, along with source data accessed by SAP’s Data Warehouse Cloud.  

Analytics would allow users to gain immediate insights into circumstances.

The combination of the smart meter data and weather forecast information would provide a calculated load profile in real-time, driving solar power production for the near future.

Digitalization pillar

The new automated system was deployed earlier in 2023, freeing IWB from ongoing manual intervention while taking an intelligent approach to forecasting the next day’s electricity production from solar PV systems.

With ML producing a total load profile in 15-minute intervals, forecasters could view solar production in eight-hour chunks.

As solar PV systems are added, the calculation methods automatically adapt.

The increased transparency allows IWB to quickly mitigate risks and develop short- and long-term countermeasures.

The solution has become a pillar of IWB’s digitization strategy, receiving international acclaim.  This year, the company was honored as a winner in the “Cutting Edge Genius” category at the SAP Innovation Awards

The annual ceremony celebrates organizations incorporating SAP technologies to effect change. You can read IWB’s pitch deck for the full details behind their amazing achievement that earned them this prestigious award.

Artificial Intelligence