Please harness the weather

HAFSLUND CASE

In district heating, the most important KPI is the weather report. Even if the forecast is reliable, the random nature of temperature fluctuations makes companies such as Hafslund Varme, part of the Oslo-based listed energy company Hafslund ASA, exposed to high and costly risks. To pick up the fight against the unpredictable Norwegian weather, Hafslund decided to bring in new competencies.

For most people, a sunny 10° celsius day in March in Oslo is good news and an encouraging sign of spring. For Hafslund Varme, it’s not. The high temperature causes households to consume one third of the energy they would have consumed on an average March day, and the revenue per energy unit is lower, as the heat price is pegged to the power price, which drops when the temperature rises. The only positive addition to the equation is that Hafslund can use less expensive boilers, reducing the average production costs. Altogether, net profits for that sunny day in March is one million Norwegian kroner or more below average results.

Bringing in a pair of fresh eyes

Nobody likes high weather-related risks, particularly not Hafslund’s shareholders: the more risk exposure, the lower the share price. This is particularly true for risks outside of the company’s control, and for more or less random factors such as wind and temperature. To reduce this vulnerability, Hafslund Varme first needed to understand exactly how weather fluctuations impact the bottom line. And to achieve this, they needed a pair of fresh eyes.

Ingeborg Sauge Torpe graduated from the University of Oslo in June 2014 with a master’s degree in Applied Mathematics. Despite no previous experience with Hafslund or district heating, Ingeborg had one thing Hafslund needed: a specialisation in numerical optimisation. Only four days after graduating, she found herself in an office at Hafslund, deeply involved in the company’s latest optimisation project. Ingeborg remembers it as quite a kick-start. “Although I had been working with optimisation in my master’s thesis, I had no other experience. This was my first job”, she says. Hafslund, however, had great expectations for Ingeborg. The project ambition was to bring all operational features together in one simulation model in order to provide strategic-level answers to the management group.

The temperature risk curve

Behind the deceptively simple issue of weather exposure, a number of complexities reside. Hafslund has 47 boilers, based on eight different fuels. These need to be switched on and off to match fluctuating demands. The costs of the fuels used in each boiler fluctuate dramatically. Most of the boilers take time to start up, require frequent service intervals and are physically located in different parts of the network, causing flexibility constraints. Also, customers have different types of contracts based on multiple fuel prices.

“It’s a bit complicated”, Ingeborg admits. But that didn’t stop her. Only three months after joining Hafslund, Ingeborg presented the “Temperature Risk Curve” to the company’s CEO demonstrating that true impact may indeed be achieved at all levels of an organisation and at all tenures. The primary purpose of the model is to provide the knowledge necessary to reducing risk exposure: By disclosing the relationship between profits and temperature, the model enables Hafslund to more efficiently hedge the risk. In addition, it may be used to optimise production planning.

Residential developments

Houses

Commercial buildings

were in 2014 connected to Hafslund Varme’s district heating grid, covering 20% of Oslo’s total heating requirements.

“With temperatures and prices constantly fluctuating – the price can change from one hour to the next – even the tiniest mistake in our systems has the potential to become a costly affair”, Ingeborg explains

From university to the CEO’s agenda in three months

Ingeborg’s Temperature Risk Curve was rapidly integrated into Hafslund Varme’s boiler operations. Today, the team applies the model on a daily basis to plan fuel combinations, determine which plants to activate when and to optimise pro casecurement and production. Meanwhile, Ingeborg continues to develop the tool for further application. “It can be used in many contexts”, she says. “For example to predict the effect of extreme scenarios. Let’s say we experience extreme weather and power cuts on the same day. Will we we still be able to deliver? I think this project has helped us become much more attentive to the choices we make every day”.

On a personal level, Ingeborg has enjoyed the ride from university to the CEO’s agenda. “I quickly got into my role while learning on the way, both about the Hafslund district heating system and about structured, hypothesis-based problem solving”. Having helped Hafslund harness the unpredictable Nordic weather appears to be merely a pleasant by-product.

Ingeborg Sauge Torpe bio:

Ingeborg joined Hafslund Varme directly from university in 2014. As a graduate in Applied Mathematics, her master’s thesis involved more partial differential equations than most people encounter in a lifetime. At the age of 27, Ingeborg is inspired by the idea of working towards sustainable and innovative energy solutions.