25 days ago
The energy market faces new, gigantic challenges

The Swedish and global energy market is facing major challenges. The increasing amount of variable renewable energy (VRE) in the market mix requires new priorities. This is because green energy sources are less predictable than fossil and nuclear energy.

It is no longer just the amount of energy that determines the economic value of an energy source. The exact timing of production and the ability to predict this timing are becoming increasingly important factors, increasing the overall value and profitability of energy production.

"VRE producers must predict their production output in order to be able to sell corresponding products, i.e. different types of futures contracts, on the markets. An error in their ability to predict their production can be very costly, since real-time transactions must compensate for the shortage or excess of production at completely different prices,” says Leonard, AI Officer at Calejo Hybrid Intelligence.

Flexible, but highly inefficient, fossil fuel-based emergency power systems must step in and a varying amount of electricity must be wasted to maintain market balance.

“Increasing shares of VRE in the market lead to a higher share of losses for each producer due to the weather-related correlation in production between producers,” continues Leonard Johard.

Hydropower still most important

In Sweden, hydropower will remain the most important source of energy storage for the foreseeable future.

“Production planning for hydropower is a complex exercise, as the value of a given potential energy content stored in water today fluctuates wildly. Storage, safety and ecological constraints limit production capacity. Complex weather patterns, high-level turbine production capacity, ground conditions, flow times, reservoir geometries and hydrological phenomena are the traditional parameters of interest in maximizing the value of water given the set of constraints,” says Leonard Johard.

Markets increasingly complex

In parallel, the increasing high-frequency fluctuations in power generation as a result of increased VRE elements have made short-term balancing of power supply more important.

“New financial products are being introduced in the power market to address these challenges and existing sales of capacity to balance power supply are becoming an important source of revenue for suppliers. Markets are introducing new short-term products at the minute level, which require modeling of new aspects such as start-up times and local flows. These new markets are not effectively handled by any of the traditional planning tools,” says Leonard Johard

It is no longer enough to balance the power market at the national or regional level.

“Weather fluctuations can create significant differences and effectively divide the market into several separate pricing areas, where equilibrium pricing must be established locally for each market. The transmission problem and the diversity of markets create additional computational demands on an already strained optimization process.”

Today's tools are insufficient

The forecasting and modeling tools available today are insufficient and strained. Updates are too infrequent to meet all new challenges.

"This leads to grossly simplified analyses and incorrect assumptions. Often, separate simple Excel sheets and various ad-hoc solutions are used to allow operators to bypass all the built-in weaknesses of the current system. Modeling the various tables and parameters in the optimization systems becomes largely a manual exercise in data collection and interpolation, which entails high costs and, due to resource limitations, much of the simulation is based on data that is several years old," says Leonard Johard.

The available optimization and analysis tools are also perceived as complex and isolated and impractical for the electricity market's business needs.

"They use pre-compiled and inaccessible code bases with a predefined but outdated usage pattern in mind. Although they are said to be critical to the business, they are not perceived as practical to use in the exact way they are intended. This often leads to further modeling errors. Power producers currently do not have any tools available to solve these problems.”

Increased costs and risks

Human traders and analysts do their best to handle planning. Larger companies therefore hire data analysts and often look for their own solutions to specific problems.

“In other words, production planning becomes more labor-intensive, as the automated tools are no longer applicable. This increases costs, as well as risks, since the essential knowledge is tied to a few key people,” says Leonard Johard.

Light in the dark

But there is light in the dark. AI company Calejo Hybrid Intelligence has begun the long-term work of creating a more intelligent tool for energy planning by municipalities, regions, states and producers.

Calejo bases its technology on hybrid intelligence – a patented mix of artificial and human intelligence. Calejo, who has long been one of the pioneers behind the concept of hybrid modeling, has over the years modeled everything from pulp mills and mining industries to power systems and robotics. Calejo is currently focusing on two areas: Smart Water and Smart Energy. Within Smart Water, successful work is already underway with several municipalities, regions and leading market players in Sweden and Italy.

Calejo Smart Energy