The balance of the electricity system requires substantial forecast data

Share
Fingrid’s operation planning unit is constantly producing forecasts of electricity consumption, generation and transmission to support the electricity system. As more and more electricity comes from weather-dependent generation – solar and wind power – the importance of forecasts will also grow.

 

Fingrid’s transmission system centre, which maintains a balanced level of power in the electricity system, plans its actions for the days ahead with the help of electricity consumption and generation forecasts for the upcoming days.

According to Mika Laatikainen, Planning Engineer in Fingrid’s operation planning unit, keeping the electricity system in balance requires forecasts to determine whether there will be enough electricity, enabling generators to react in good time in the event of a shortfall. The transmission forecast is also important.

“The transmission forecast tells us how much electricity will be consumed and generated in various areas and whether we will be able to transfer the necessary electricity between these areas,” he says.

Consumption and generation forecasts are required to provide this information.

The generation forecast is based on the generation plans available from power plants that utilise conventional forms of generation, as well as the forecasts of weather-dependent wind and solar electricity generation.

Wind power forecast constantly updated

The wind power forecast has grown in importance in recent years as wind farms have been built at a frenetic rate. The most important piece of information for forecasting wind power generation is the wind speed.

“The forecasting model contains the coordinates and capacity of every wind farm in Finland. The software uses the coordinates to fetch a weather forecast that is as localised as possible to the wind farm. By comparing the weather forecast with the actual generation figures under equivalent conditions, we are usually able to forecast the output very accurately,” Laatikainen says.

The wind forecast is constantly updated as the wind farms provide Fingrid with their actual generation measurements.

“This helps us to monitor the accuracy of our forecast and improve the model,” Laatikainen says.

Solar power forecast based on capacity and radiation forecasts

Similar reference data is not yet available to support solar power generation forecasts, as the majority of Finland’s solar power systems are connected to domestic electricity meters.

The solar power forecast is based on data provided by the Energy Authority to Fingrid concerning the solar power capacity in the various areas covered by each distribution system operator.

“The forecast is created by combining the capacity data with the hourly forecasts of solar radiation for the specific coordinates.”

According to Laatikainen, Fingrid aims to obtain data on the actual amount of solar power generated, as this will help it to develop its forecasts.

“For example, in Denmark, suppliers of solar power systems provide the transmission system operator with anonymised data on the total amount of electricity generated. This could also be possible in Finland. However, solar power currently accounts for such a small proportion of total generation that we have not deemed this necessary,” Laatikainen says.

Consumption forecasts based on years of data and weather forecasts

Laatikainen says that electricity consumption forecasts have been prepared for a long time, and the forecasts, which are based on data accumulated over several years, are generally reliable.

“The weather forecast is perhaps the most significant individual – and constantly varying – factor in the consumption forecast. The temperature, wind speed and wind chill factor determine how much electricity will be consumed for purposes such as domestic heating.

Other variables that the forecast has learned to take into automatic consideration include industrial loads, which are not dependent on weather conditions, and the effects of weekdays, weekends and holidays.

No manual work is involved in preparing daily forecasts. The forecasts are produced automatically by software that fetches the necessary data from various sources. In addition to its own forecasts, Fingrid also purchases ready-made forecasts.

“It is our job to make sure that the automatic systems that generate forecasts keep running. We also endeavour to improve the accuracy of forecasts over the long term.”

Chilling effects on wind power forecasts

The amount of wind power generated in January 2019 fell well below Fingrid’s forecast. The impact of freezing temperatures contributed to the lower output. Fingrid now aims to develop a forecast that takes freezing into account.

Freezing refers to the ice that forms on the blades of wind turbines in winter conditions. The layer of ice alters the aerodynamic properties of the blades, preventing the turbine from generating at full capacity. Although Fingrid’s personnel were familiar with the concept of freezing turbines, the impact on wind turbine output only became fully apparent in January 2019.

“On 10 January, wind power generation fell markedly below our forecasts. The shortfall was made up with purchases from the balancing power markets,” says Mika Laatikainen, who develops forecasts in his role as a Planning Engineer at Fingrid.

As Fingrid considered the reasons for the forecasting error, freezing stood out as one of the credible alternatives. The multidisciplinary team working on the BCDC Energy research project were also interested in the case. In a related blog post, BCDC’s Anders Lindfors and Santtu Karhinen found in their analysis that freezing had played a major role in the forecasting error, and they emphasised that electricity markets should be more aware of the impact of freezing.

“At the moment, the impact of freezing is not incorporated in Fingrid’s wind power forecasts. In the wake of the case in January, an EU-funded project is underway in which the Finnish Meteorological Institute will develop a wind power estimate that takes the effect of freezing into account. We are involved in the project as an end-user,” Laatikainen says.

Comments

Leave a Reply

Your email address will not be published.