Storage Scheduling with Stochastic Uncertainties: Feasibility and Cost of Imbalances

TL;DR

This paper introduces a novel probabilistic scheduling method for energy storage systems that manages uncertainties in renewable energy and loads, ensuring feasible dispatch with a specified probability, and compares it to scenario optimization.

Contribution

It proposes a new probabilistic dispatch scheduling approach using cumulative density functions, enhancing feasibility guarantees under uncertainty, and compares it with existing scenario-based methods.

Findings

The probabilistic method ensures dispatch feasibility with a pre-set probability.

Scenario optimization focuses on minimizing expected costs, but may lack feasibility guarantees.

Simulations demonstrate the effectiveness of the proposed approach over traditional methods.

Abstract

Dispatchability of renewable energy sources and inflexible loads can be achieved using a volatility-compensating energy storage. However, as the future power outputs of the inflexible devices are uncertain, the computation of a dispatch schedule for such aggregated systems is non-trivial. In the present paper, we propose a novel scheduling method that enforces the feasibility of the dispatch schedule with a pre-determined probability based on a description of the operation of the system as a two-stage decision process. Thereby, a crucial point is the use of probabilistic forecasts, in terms of cumulative density function, of the inflexible energy consumption/production profile. Then, for the sake of comparison, we introduce a second scheduling method based on state-of-the-art scenario optimization, where, unlike the proposed method, the focus is on the minimization of the expected final cost. We draw upon simulations based on real consumption and production data to compare the methods and illustrate our findings.