Load Forecasting Attacks

Code repo for ACM e-Energy 2019 paper: Exploiting Vulnerabilities of Load Forecasting Through Adversarial Attacks Paper [Runner-Up for Best Paper]

Extended version: Vulnerabilities of Power System Operations to Load Forecasting Data Injection Attacks

Authors: Yize Chen, Yushi Tan, Ling Zhang and Baosen Zhang

University of Washington

Contact: yizechen@uw.edu

Introduction

Load forecasting plays a critical role in the operation and planning of power systems. By using input features such as historical loads and weather forecasts, system operators and utilities build forecast models to guide decision making in commitment and dispatch. As the forecasting techniques becomes more sophisticated, however, they also become more vulnerable to cybersecurity threats. We study the vulnerability of a class of load forecasting algorithms and analyze the potential impact on the power system operations, such as load shedding and increased dispatch costs. Specifically, we propose data injection attack algorithms that require minimal assumptions on the ability of the adversary. By only injecting malicious data in temperature from online weather forecast APIs, an attacker could manipulate load forecasts in arbitrary directions and cause significant and targeted damages to system operations.

Attack Method

Learn and Attack

Gradient Estimation Attack

Load Forecasting Algorithms

Code Dependencies

All code are implemented in Python.

Learning algorithms for load forecasting: Tensorflow and Keras

Power systems unit commitment and economic dispatch: Pypsa

Run Experiments with PreProcessed Datasets

We make use of

weather data from Dark Sky API;

load data from ENTSOE(European Network of Transmission System Operators for Electricity).

June 17th, 2019: Updates on attacking a power network

Vulnerabilities of Power System Operations to Load Forecasting Data Injection Attacks Paper