Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

DiffLoad: Uncertainty Quantification in Electrical Load Forecasting with the Diffusion Model

Zhixian Wang, Qingsong Wen, Chaoli Zhang, Liang Sun, Yi Wang

TL;DR

This paper proposes a diffusion-based Seq2seq structure to estimate epistemic uncertainty and employs the robust additive Cauchy distribution to estimate aleatoric uncertainty and demonstrates the ability to separate and model the two types of uncertainties for different levels of loads.

Abstract

Electrical load forecasting plays a crucial role in decision-making for power systems, including unit commitment and economic dispatch. The integration of renewable energy sources and the occurrence of external events, such as the COVID-19 pandemic, have rapidly increased uncertainties in load forecasting. The uncertainties in load forecasting can be divided into two types: epistemic uncertainty and aleatoric uncertainty. Separating these types of uncertainties can help decision-makers better understand where and to what extent the uncertainty is, thereby enhancing their confidence in the following decision-making. This paper proposes a diffusion-based Seq2Seq structure to estimate epistemic uncertainty and employs the robust additive Cauchy distribution to estimate aleatoric uncertainty. Our method not only ensures the accuracy of load forecasting but also demonstrates the ability to separate the two types of uncertainties and be applicable to different levels of loads. The relevant code can be found at \url{https://anonymous.4open.science/r/DiffLoad-4714/}.

DiffLoad: Uncertainty Quantification in Electrical Load Forecasting with the Diffusion Model

TL;DR

This paper proposes a diffusion-based Seq2seq structure to estimate epistemic uncertainty and employs the robust additive Cauchy distribution to estimate aleatoric uncertainty and demonstrates the ability to separate and model the two types of uncertainties for different levels of loads.

Abstract

Electrical load forecasting plays a crucial role in decision-making for power systems, including unit commitment and economic dispatch. The integration of renewable energy sources and the occurrence of external events, such as the COVID-19 pandemic, have rapidly increased uncertainties in load forecasting. The uncertainties in load forecasting can be divided into two types: epistemic uncertainty and aleatoric uncertainty. Separating these types of uncertainties can help decision-makers better understand where and to what extent the uncertainty is, thereby enhancing their confidence in the following decision-making. This paper proposes a diffusion-based Seq2Seq structure to estimate epistemic uncertainty and employs the robust additive Cauchy distribution to estimate aleatoric uncertainty. Our method not only ensures the accuracy of load forecasting but also demonstrates the ability to separate the two types of uncertainties and be applicable to different levels of loads. The relevant code can be found at \url{https://anonymous.4open.science/r/DiffLoad-4714/}.
Paper Structure (19 sections, 3 theorems, 26 equations, 6 figures, 7 tables, 2 algorithms)

This paper contains 19 sections, 3 theorems, 26 equations, 6 figures, 7 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Background and Motivation
  3. Literautre Review
  4. Contributions
  5. Paper Organization
  6. Proposed Method
  7. Framework
  8. Epistemic Uncertainty Quantification
  9. Aleatoric Uncertainty Quantification
  10. Training and Inference
  11. Case Studies
  12. Experimental Setups
  13. Experimental Results
  14. Results Analysis and Discussion
  15. Epistemic Uncertainty Estimation
  16. ...and 4 more sections

Key Result

Lemma 1

cheng2024robusttsf Let $\ell$ be the loss function, $f$ be the forecasting model, $R_{\ell}(f)$ be the empirical loss on the clean training set, and $R^\eta_{\ell}(f)$ be the empirical loss on the training set with noise. Under different noise anomalies with anomaly rate $\eta<0.5$, we have

Figures (6)

  • Figure 1: Visualization of sudden change effects caused by external events in electricity load
  • Figure 2: Overview of our proposed DiffLoad method.
  • Figure 3: Comparison of metrics on ten building datasets.
  • Figure 4: Visualization of 75 $\%$ interval of two datasets.
  • Figure 5: Epistemic uncertainty estimation on COV dataset between different methods
  • ...and 1 more figures

Theorems & Definitions (5)

  • Lemma 1
  • Theorem 1
  • proof
  • Definition 1
  • Lemma 2