Evaluating the sensitivity of the total system cost and installed capacity of technologies of the electricity system to the perturbation of the wind and PV capacity factor

Increasing renewable energy penetration is essential for achieving carbon neutrality in the electricity system. In this regard, the most promising technologies are wind and PV. The degree of penetration of these technologies in the mix is affected by their capacity factor.

The objective of this study is to determine the sensitivity of the electricity system to changes in the capacity factor of wind and PV, not only uniform changes but also the changes in the low and high wind or PV production conditions. Simulations were performed using EOLES, an investment and dispatch optimization model. This model minimizes the total system cost by satisfying hourly demand, respecting technical and operational constraints, and giving us the optimal electricity system for a given input. This result provides an overview for the decision makers deciding how much capacity to install. In addition, to reflect the realistic situation of the energy system, in which we have already invested in installed capacities, EOLES is used only for dispatch optimization with the pre-fixed installed capacities.

Output variables chosen for sensitivity tests are total system cost and installed capacity of production technologies. Their sensitivity to changes in the average capacity factor was measured using elasticity quantity, which is calculated by dividing the relative change of the chosen output variable by the relative change of the capacity factor average. Uncertainty of capacity factor in the different production conditions of wind and PV was modeled by perturbing a specific quantile of the capacity factor dataset at each test and uniform errors by uniform perturbation of all time steps. Furthermore, perturbations of different magnitudes and signs are included to show the behavior of EOLES concerning the amount of perturbation.

The result shows the EOLES model is more sensitive to change in the capacity factor of the wind and least to PV for both Installed capacities and total system cost; also, it is more sensitive to the perturbation of low-production than high-production conditions. For instance, the elasticity of the installed capacity of PV and wind to perturbation of their capacity factor in low-production conditions is 15 and one, respectively, and it is approximately zero for both PV and wind in high-production conditions.

Optimization of installed capacities and dispatch in response to capacity factor perturbations results in a weak sensitivity of the total system cost (elasticities less than 0.5). On the other hand, optimizing only dispatch leads to having the elasticity of the total system cost as high as 14. Comparing elasticities indicates that installed capacity optimization compensates for the effect of capacity factor perturbation on total system cost. However, fixed installed capacity leads to either having an oversize system in positive or extra usage of expensive reserve technologies in negative perturbations; as a result, the higher elasticity of the total system is expected. Considering the high sensitivity of the low production events of the wind, it is worth improving our modeling of smaller capacity factors, including choosing a wind dataset, a bias correction method, and a power curve.