Heat demand model
The general approach for heat demand calculation is to determine the building archetypes (following the method from the EU TABULA project). For each country, a set of categories of common building types are defined. These are linked to typical energy use patterns. We derive energy use by typology from measured data where possible.
Swiss heat demand model
The basic Swiss heat demand model is based on the work of Schneider et al. 2019.
For heating and domestic hot water it calculates:
- yearly useful energy demand
- yearly final energy demand
- heating system efficiency
- heating system fuel CO₂ emissions intensity
- yearly CO₂ emissions
- hourly load curves
The yearly demands are based on climate-corrected energy demand per m² for building archetypes defined by building type (single family, multifamily, office, etc.) and construction year. Energy demand intensities are mainly based on measured data from Geneva (IDC data).
An important aspect of the model is the application of multiple levels of data filling and assumptions as a function of the completeness of the input data. Wherever possible, values from the input data are used; where they are missing, they are replaced by models and assumptions. Extra columns with names following the pattern *_origin track where the values in the final results come from.
The load curves are based on measured load curves by archetypes with a method for introducing hourly variability on load curve aggregates. For each load curve typology there is a family of demand deviations (or perturbations). The typical daily demand is calculated using a mean daily temperature regression method, then the hourly perturbations are sampled from the load curve family.
The hourly load curve model can run independently from the yearly model, as it performs the perturbation step on a normalised basis and then scales for yearly demand, so yearly demand from other sources can be used (e.g. measurements).
The class SwissHeatDemandModel stores the parameters necessary for running the model and allows the parameters to be provided as arguments to override built-in values.
Future energy demand
To calculate future heat demand predictions, we determine the potential energy saving from renovation for each building archetype. Whether a building is renovated or not can either be set by the user, or we can apply a machine learning (probability) model that estimates when a building is likely to be renovated.