Denoising Autoencoder-Based Missing Value Imputation for Smart Meters

Abstract

Electric load data are essential for data-driven approaches (including deep learning) in smart grid, and advanced smart meter technologies provide fine-grained data with reliable communications. Despite the recent development of smart metering devices, however, missing data still arise due to unexpected device power off, communication failure, measuring error, or other unknown reasons. In this paper, we investigate a deep learning framework for missing imputation of smart meter data by leveraging a denoising autoencoder (DAE). Then, we compare the performance of the proposed DAE with traditional methods as well as other recently developed generative models, e.g., variational autoencoder and Wasserstein autoencoder. The proposed DAE based imputation shows significantly better results compared to other methods in terms of root mean square error (RMSE) by up to 28.9% for point-wise error, and by up to 56% for daily-accumulated error. INDEX TERMS Deep learning, smart grid, missing imputation, smart meters, denoising autoencoder, generative model, daily load profile (DLP).

Supervised Learning Approach for State Estimation of Unmeasured Points of Distribution Network

“…. Equations (6) and (7) are the maximum absolute magnitude error of the estimated voltage magnitude and the angle for the m th of Mtar target consumers, respectively. The elements of the matrices in (6) and 7are described in (8) and (9).…”

“…Aggregators are required to manage vast amount of data, which also incurs costs. In addition, data missing still occurs due to unexpected device power off, communication failure, measurement error, or other unknown errors [7]. It is still not clear how the communications infrastructure for transmitting expected measurement data to a distribution management system (DMS) will change [6].…”

Three‐step imputation of missing values in condition monitoring datasets

“…(iii) In recent years, different types of deep neural networks (DNNs) are applied to establish spatial-temporal-based models to impute missing data in condition monitoring datasets. For instance, autoencoder (AE) and its variants were constructed to repair DGA data, load data, and wind turbine condition parameters [21][22][23]. In addition, multilayer perceptron, generative adversarial networks, and recurrent neural networks were also widely applied [24][25][26].…”

“…According to Algorithm 1, the continuous missing values of OilT and C 2 H 6 should be imputed in sequence. The prediction functions of OilT and C 2 H 6 are shown in (21) and (22), respectively…”

Missing-Insensitive Short-Term Load Forecasting Leveraging Autoencoder and LSTM

“…The methods of missing imputation can be categorized by a) linear interpolation, b) historical average, c) deep learningbased [16]. First, the linear interpolation method replaces missing values with an average of measured values, which occur before and after missing.…”