Survivability Analysis of Local Energy Systems under Critical Supply Faults : Quantifying the Role of Wind Power and Storage in Energy Resilience

Sammanfattning

When a winter grid outage disconnects a residential district from both its grid electricity and its CHP-fed district heating, backup heat pumps become the sole heat source, powered only by a local wind farm and battery storage. The same fault that removes the heat supply also removes the means to recharge the battery from the grid. This degree project quantifies how long 1070 dwellings in Eskilstuna remain warm under that condition, and whether a centralised or distributed battery architecture survives longer. Stochastic bottom-up electrical demand profiles were generated with RAMP and calibrated to Swedish Energy Agency end-use statistics. Thermal demand was modelled independently following ISO 52016-1, parameterised on measured Swedish building data. A wind time series for two Enercon E-82/2300E2 turbines was derived from ERA5 reanalysis using PyWake wake modelling. Battery capacities were sized by mixed-integer linear programming peak-shaving optimisation. Survivability was measured in dwelling-hours, defined as the number of dwellings staying above 18°C summed over time. Under the worst-case winter outage, the centralised architecture (Scenario A) lasted 22h and the distributed architecture (Scenario B) lasted 16h. A 12h pre-charge warning extended survival by 32h in Scenario A (to 54h) and by 12h in Scenario B (to 28h), an asymmetry that follows from the pooling disadvantage of four independent cluster batteries against one centralised unit. Both gains exceed the 6h gap between architectures. Under component failure, Scenario B degraded more gradually than Scenario A. The main finding is that issuing a 12h warning matters more than which architecture is chosen. It more than doubles Scenario A survival time and requires only forecast issuance rather than added battery capacity.