An Improved Load Forecasting Method Based on the Transfer Learning Structure under Cyber-Threat Condition

Abstract

Smart grid is regarded as an evolutionary regime of existing power grids. It integrates artificial intelligence and communication technologies to fundamentally improve the efficiency and reliability of power systems. One serious challenge for the smart grid is its vulnerability to cyber threats. In the event of a cyber attack, grid data may be missing; subsequently, load forecast and power planning that rely on these data cannot be processed by generation centers. To address this issue, this paper proposes a transfer learning-based framework for smart grid scheduling that is less reliant on local data while capable of delivering schedules with low operating cost. Specifically, the proposed framework contains (1) a power forecasting model based on transfer learning which can provide high quality load prediction with limited training data, (2) a novel adaptive time series prediction method with modeling time series from a covariate shift perspective that aims to train the forecasting model with a strong generalization capability, and (3) a day-ahead optimal economic power scheduling model considering a shared energy storage station.

A Review of AI-Based Cyber-Attack Detection and Mitigation in Microgrids

“…In smartgrids, cyber-attacks may impede access to local data which can cause issues in power planning and dispatch decisions. Deep transfer learning for load forecasting can provide high-quality load prediction with less data so that in case of missing local data the prediction data are readily available [116]. In general, forecasting methods can be improved by utilizing the generalizing capability of transfer learning without the need for excessive data.…”

Modeling and predictive control of nonlinear processes using transfer learning method

“…Additionally, transfer learning has received increasing attention in process systems engineering to address a variety of problems. [14][15][16][17][18] For example, transfer learning was adopted to design a fault detection and diagnosis method for multimode chemical processes in Wu and Zhao. 16 A process monitoring method with high generalizability was developed using transfer learning in Zhu et al 18 A new transfer learning strategy was proposed to solve the problem of end-product quality for new batch processes in Jia et al 14 Additionally, transfer learning was used to improve the simulation accuracy for material property prediction in Li et al 15 Despite the increasing popularity and application of transfer learning in many fields of process systems engineering, the application of transfer learning in process modeling is still in its infancy.…”

“…For example, in Tzeng et al, 13 an adaptation layer was introduced to a convolutional neural network (CNN) to adapt the pretrained CNN to the target domain with fine‐tuning, for which the modeling accuracy was significantly improved. Additionally, transfer learning has received increasing attention in process systems engineering to address a variety of problems 14–18 . For example, transfer learning was adopted to design a fault detection and diagnosis method for multimode chemical processes in Wu and Zhao 16 .…”