The hidden cost of using time series aggregation for modeling low-carbon industrial energy systems: An investors’ perspective

Time series aggregation (TSA) is commonly used in energy system optimization to reduce model complexity and computational expenses by selecting periods to represent the entire time series. TSA’s accuracy has traditionally been assessed by comparing the objective values between the original and TSA models (assumed error). However, evaluating TSA from an investor’s standpoint involves analyzing the performance of TSA-based energy system designs using the original time series. Therefore, we introduce the hidden error and total error, novel error metrics, to evaluate the financial implications of TSA through backtesting the TSA-based system designs using the original time series. Our analysis extends to the effects of carbon removal prices and the choice between marginal and grid-mix emission factors on TSA’s performance. We find that traditional error metrics inadequately capture TSA-induced financial losses. We demonstrate that for ambitious emission reduction targets, the hidden error, which had not been identified previously, is up to 29 times the error identified by prior error evaluation approaches. Consequently, TSA should be applied cautiously, particularly when outcomes are vital for investment decisions and when lower TSA rates yield computationally feasible models, highlighting the need to consider comprehensive error metrics in TSA applications for more accurate energy system optimization.