Cutting down unnecessary initiation of power plants and excessive district heating water temperature with hyperlocal weather forecasts is a new way forward.
By Tuukka Teppola, Growth Manager – Wx Beacon Vaisala Xweather
Published in Hot Cool, edition no. 3/2023 | ISSN 0904 9681 |
Every district heating company faces this challenge: there is an upcoming cold weather front, and as a heating provider, you strive to meet the demand. You must decide how much heat to produce and adjust your network’s water temperature. It’s cold outside, so you initiate your additional heating plant and start burning fuel. You raise the water temperature and get ready to pump the water to every corner of the network. But the cold weather never comes, and you end up with increased emissions and heat loss, all for nothing. It would help if you had better weather forecasting in your planning, one that can account for your city’s unique characteristics and minimizes the errors in the forecast.
Minimizing unnecessary plant initiations and optimizing water temperature are feasible and relatively easy ways to reduce environmental footprint (and potentially increase profits). Initiating and running additional plants consumes enormous amounts of energy and other resources. The same is true for the excessively high-water temperature in the network: extra fuel, unnecessary production mix, electricity for pumping, person-hours, heat losses in pipes; you name it.
When trying to choose the perfect mix of optimal heat production, the right supply temperature, and the correct heat distribution, heat load forecast is everything. Accurate heat load forecasting helps district heating companies operate more efficiently and effectively, improving customer satisfaction, lowering environmental burdens, and increasing profits.
Two key factors that impact load forecast are human behaviour and weather. While human behaviour is quite predictable — based on daily, weekly, and annual patterns — weather prediction is a much trickier case. That is why district heating companies depend heavily on weather forecasts. Unfortunately, significant forecast errors are not uncommon; plants then waste fuel and increase emissions. It all comes down to correctly forecasting the weather in the location that matters for your energy operations — for upcoming minutes, hours, and days.
Now, regarding weather forecasts, the number of errors is everything.
Accurately representing the current state of the atmosphere begins with observations. Today, we leverage both in-situ (in place) and remote sensing networks to observe the atmosphere. In-situ, surface-based networks are numerous but only represent a small part of the atmosphere.
Remote sensing capabilities, primarily from space, measure more significant portions of the atmosphere but struggle to see all variables (especially near the surface) and have their own challenges (like getting satellites to space and keeping them functioning there).
Generally, surface-based networks are built to serve the general public and transportation infrastructure (e.g., airports and harbours). For this reason, they do not capture all the local weather patterns that are levant to a particular district heating network. Yet, understanding those patterns is the key to efficient weather forecasting in the context of district heating operations. Even within a mid-sized city, outdoor temperatures can vary drastically from one location to another.
A city centre with a densely built environment can, for example, have a different microclimate compared to a nearby district located either close to a body of water, higher above the sea level, or in a valley. The temperature difference within just several hundred meters can vary by several degrees. This is where accurate, hyperlocal weather forecasts can help. A data-driven heat production plan starts with a correct prediction.
Wx Beacon by Vasiala is an enhanced hyperlocal weather forecast that measures local conditions in the areas of customer interest to ensure the best possible accuracy. It carefully considers important local topography factors such as building environment, water systems, and vegetation in various parts of the district heating network.
Local measurements (space-proof technology) are combined with an in-house forecasting model (top-ranked globally), using AI/ML technologies, improving the city’s regional accuracy of the weather forecast. “Let’s look at the example of Fortum’s network in Espoo, Finland.
Due to microclimates within the city, air temperatures can vary significantly, even over short distances. The graph demonstrates air temperature data from two hyperlocal identical weather sensors only 7 km apart in the city of Espoo, Finland
The graph above demonstrates how significantly temperature can vary between measurement points within a relatively small area. Adding hyperlocal Wx Beacon forecast enhanced with sensor observations decreased the number of significant errors (over 2.5°C) by 74% and improved overall accuracy by up to 36%, helping minimize heat loss and unnecessary emissions.
With regards to CO2, monetary, and resource-saving, every network differs. To contextualize this, we can distinguish several different aspects in which more accurate local weather can benefit DHC companies:
- Help to avoid initiating fossil plants and instead operate with a greener heat portfolio.
- Allow lower water temperature. For example, a 10°C decrease in water temperature is estimated to lead to 8.5% reduced heat losses*.
- Minimize situations where CHP is driven by the electricity-price-first approach, not the heat-demand-first approach (heat to scrap).
- Help to make better decisions on spot markets. With CHP peak electricity production of 50MW, estimation of 150k€ yearly savings.
- Save electricity thanks to optimal water pumping.
*2021, Ikävalko, Master’s Thesis
For further information please contact: Tuukka Teppola, tuukka.teppola@vaisala.com
“Hyperlocal weather forecast” was published in Hot Cool, edition no. 3/2023
