Combined Heat and Power Model for the Nordic Region Released

Oct. 16, 2017Katinka Bogaard

A new model with an hourly resolution has been developed for combined heat and power (CHP) production in the Nordic region, more specifically each individual zone in Sweden, Denmark and Finland, as well as for aggregate Norway.

Model Description

Through the use of multiple regression analysis, parameters have been estimated for numerous variables to fit actual hourly data covering the time span January 2014 to January 2017. The model takes into account fluctuations in temperature, price and consumption, while also including hourly, monthly and seasonal variations.

Further, each model is adjusted to coincide with statistical values as reported by respective TSOs and statistical bureaus for each country.

Model Performance

An example of model results for a given winter and summer month, respectively, in Finland is given in the graphs below.

Combined Heat and Power Model for the Nordic Region Released
Figure 1: Model Performance for a Given Winter Month in FinlandCombined Heat and Power Model for the Nordic Region Released
Figure 2: Model Performance for a Given Summer Month in Finland

It has been observed that the CHP production in the different regions is driven by different parameters.  The models for the regions are therefore of varying quality and precision. However, all models are able to capture the majority of variation and prove to be highly correlated with the actual production.

Implementation and Further Work

This model is to be included in the instant spot model,  giving a more complete input and thus more representative results for the Nordic region. The output from this model can be found under daily CHP production values on the Instant Spot Overview page. For a more detailed overview, we have also published hourly forecast and history overview pages. A change log and further updates can be found here.

Resolution and Areas

It is important to note that this is a preliminary presentation of the model, and a higher resolution is likely to come in the future.  Further, the model will be expanded to cover other areas, such as the CWE area. However, due to varying data availability and quality, it is unknown when these models will be completed.