Improved super-resolution perception convolutional neural network for photovoltaics missing data recovery

Article Engineering, Civil

Deep learning based missing data recovery of non-stationary wind velocity

Qiushuang Lin et al.

Summary: This study proposes a signal-based method to recover missing non-stationary wind velocity data in wireless sensor network caused by random data loss during data transmission. The method decomposes the data using wavelet transform and empirical wavelet transform, performs preliminary recovery using convolutional neural network on high-frequency sub-signals, and then corrects the results using Kriging interpolation for extreme value estimation, achieving good recovery performance.

JOURNAL OF WIND ENGINEERING AND INDUSTRIAL AERODYNAMICS (2022)

Add to Collection

Article Green & Sustainable Science & Technology

Impact of duration and missing data on the long-term photovoltaic degradation rate estimation

Irene Romero-Fiances et al.

Summary: Accurate quantification of photovoltaic system degradation rate is essential for lifetime yield predictions. Different techniques applied to synthetic PV system data showed varying levels of accuracy, affected by evaluation duration and missing data. Filtering out corrupted data and applying a change-point detection stage are necessary for accurate estimation of degradation rate.

RENEWABLE ENERGY (2022)

Add to Collection

Article Green & Sustainable Science & Technology

Analysis of data errors in the solar photovoltaic monitoring system database: An overview of nationwide power plants in Korea

Jongwoo Choi et al.

Summary: This study provides an error case report derived from an analysis of nearly 2000 solar photovoltaic power plants distributed throughout Korea. The impact of errors on data-driven PV analysis is evaluated, with a focus on the effect of errors in system specification data on day-ahead hourly PV power forecast. The study finds that errors in system specification data significantly decrease the forecast accuracy, and regression imputation in the time domain is an acceptable method for error handling.

RENEWABLE & SUSTAINABLE ENERGY REVIEWS (2022)

Add to Collection

Article Green & Sustainable Science & Technology

A short-term wind power prediction model based on CEEMD and WOA-KELM

Yunfei Ding et al.

Summary: This study proposes a hybrid forecasting model based on CEEMD and WOA-KELM for short-term wind power prediction. The non-stationary wind power time series is decomposed into relatively stationary components by CEEMD, and these components are used as the training set for the KELM prediction model. Experimental results show that the proposed method performs well in terms of accuracy, stability, and computational costs.

RENEWABLE ENERGY (2022)

Add to Collection

Article Computer Science, Artificial Intelligence

Missing value estimation using clustering and deep learning within multiple imputation framework

Manar D. Samad et al.

Summary: This paper proposes methods to improve the imputation accuracy of the MICE algorithm by using ensemble learning and deep neural networks. The results of extensive analyses on multiple datasets show that the proposed methods outperform other state-of-the-art imputation algorithms, leading to better imputation accuracy and classification accuracy.

KNOWLEDGE-BASED SYSTEMS (2022)

Add to Collection

Article Energy & Fuels

Missing wind speed data reconstruction with improved context encoder network

Bo Jing et al.

Summary: This paper proposes an improved context encoder network (ICE) for reconstructing missing wind speed data. By using a joint loss function and a well-designed neural network architecture, the stable and near-real reconstructed wind speed data is obtained, and its effectiveness is demonstrated through experiments.

ENERGY REPORTS (2022)

Add to Collection

Article Engineering, Electrical & Electronic

Enhancing the Spatio-Temporal Observability of Grid-Edge Resources in Distribution Grids

Shanny Lin et al.

Summary: This paper presents a joint recovery framework that enhances the observability of residential loads by integrating smart meter data and phasor measurement data, effectively addressing the lack of data issue. The proposed method successfully identifies EV charging events and infers the total PV output.

IEEE TRANSACTIONS ON SMART GRID (2021)

Add to Collection

Article Computer Science, Information Systems

An intelligent scheme for big data recovery in Internet of Things based on Multi-Attribute assistance and Extremely randomized trees

Hongju Cheng et al.

Summary: Due to the constraints of sensor nodes in the Internet of Things, data loss is a key issue for applications dependent on data completeness. Current solutions, such as interpolation, use data correlation, but it can be difficult to extract data correlation in cases where the coupling degree between different attributes is low. This paper proposes an intelligent recovery scheme based on Multi-Attribute assistance and Extremely randomized Trees (MAET) for big data in the Internet of Things, which has shown efficiency and improved accuracy compared to other algorithms.

INFORMATION SCIENCES (2021)

Add to Collection

Article Green & Sustainable Science & Technology

PV Generation Forecasting With Missing Input Data: A Super-Resolution Perception Approach

Wei Liu et al.

Summary: A new forecasting framework using SRPCNN and SCN is proposed to address missing data in PV generation forecasting. Testing results show excellent missing-data recovery ability and accurate PV generation forecasting even under severe missing-data conditions. By comparing with existing methods, the proposed method demonstrates promising performance.

IEEE TRANSACTIONS ON SUSTAINABLE ENERGY (2021)

Add to Collection

Article Green & Sustainable Science & Technology

Driving force model to evaluate China's photovoltaic industry: Historical and future trends

Zimeng Luo et al.

Summary: Depletion of traditional energy sources and environmental pollution hinder meeting global energy demands. Solar PV technology has been widely used in China to replace traditional energy. Factors affecting the PV industry's development in China are complex, with policy being crucial in the past and technology and demand expected to be more significant in the future.

JOURNAL OF CLEANER PRODUCTION (2021)

Add to Collection

Article Automation & Control Systems