Condition monitoring of wind turbines with the implementation of spatio-temporal graph neural network

Liu, Jiayang; Wang, Xiaosun; Xie, Fuqi; Wu, Shijing; Deng, Li

doi:10.1016/j.engappai.2023.106000

Cited by 42 publications

(6 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

Section: F Automl E Transferência De Aprendizadounclassified

“…O meta-aprendizado é um dos tópicos relacionados ao AutoML e refere-se a capacidade do sistema aprender a aprender [43]. Nesse sentido, o objetivo do meta-aprendizado é reaproveitar as experiências prévias em tarefas que serão realizadas futuramente [7], [43]. Dessa forma, o sistema não precisa realizar o aprendizado do zero e pode adaptar-se a situac ¸ão atual a partir das suas experiências prévias [44].…”

Section: F Automl E Transferência De Aprendizadounclassified

See 1 more Smart Citation

AutoML e transferência de aprendizado por reforço entre problemas de otimização combinatória

Barbosa Souza,

Carvalho Ottoni

2024

Anais Do XVI Congresso Brasileiro De Inteligência Computacional

View full text Add to dashboard Cite

O Aprendizado por Reforço (AR) é uma das linhas do Aprendizado de Máquina e que pode ser aplicado às mais diversas situações. A Otimização Combinatória, por exemplo, é uma relevante área de aplicação do AR. Destaca-se que, na literatura recente, tem ocorrido uma alta movimentação para realizar o desenvolvimento de sistemas de AR para resolução de problemas conheci- dos de Otimização Combinatória, como o Problema do Caixeiro Viajante (PCV) e o Sequential Ordering Problem (SOP). No entanto, um desafio em aberto nessa área é a automatização da transferência de aprendizado entre o PCV e o SOP. Neste sentido, o objetivo deste trabalho foi propor um sistema de transferência de aprendizado com AutoML para otimização do SOP a partir de bases de conhecimento geradas com instancias assimétricas do PCV. Essa abordagem foi denominada de Auto TL RL. Os resultados obtidos pelo Auto TL RL mostram que, para a maioria das instâncias analisadas (aproximadamente 78%), o sistema proposto alcançou os melhores resultados em comparação aos resultados obtidos sem transferência aprendizado.

show abstract

Section: F Automl E Transferência De Aprendizadounclassified

AutoML e transferência de aprendizado por reforço entre problemas de otimização combinatória

Barbosa Souza,

Carvalho Ottoni

2024

Anais Do XVI Congresso Brasileiro De Inteligência Computacional

View full text Add to dashboard Cite

show abstract

“…An et al [32] used GNNs to learn the structure and semantic knowledge of sentence relationships for sentiment analysis. Liu et al [33] monitored the state of machinery through spatiotemporal neural networks, thereby discov-ering abnormal operations and ensuring the long-term stable operation of the machinery. Deshpande et al [34] proposed a quantum optimization algorithm based on graph neural networks that uses performance prediction to choose between quantum optimizers and classical optimizers to solve the grid parameter optimization problem.…”

Section: Graph Neural Networkmentioning

confidence: 99%

Multi-Order-Content-Based Adaptive Graph Attention Network for Graph Node Classification

et al. 2023

View full text Add to dashboard Cite

In graph-structured data, the node content contains rich information. Therefore, how to effectively utilize the content is crucial to improve the performance of graph convolutional networks (GCNs) on various analytical tasks. However, current GCNs do not fully utilize the content, especially multi-order content. For example, graph attention networks (GATs) only focus on low-order content, while high-order content is completely ignored. To address this issue, we propose a novel graph attention network with adaptability that could fully utilize the features of multi-order content. Its core idea has the following novelties: First, we constructed a high-order content attention mechanism that could focus on high-order content to evaluate attention weights. Second, we propose a multi-order content attention mechanism that can fully utilize multi-order content, i.e., it combines the attention mechanisms of high- and low-order content. Furthermore, the mechanism has adaptability, i.e., it can perform a good trade-off between high- and low-order content according to the task requirements. Lastly, we applied this mechanism to constructing a graph attention network with structural symmetry. This mechanism could more reasonably evaluate the attention weights between nodes, thereby improving the convergence of the network. In addition, we conducted experiments on multiple datasets and compared the proposed model with state-of-the-art models in multiple dimensions. The results validate the feasibility and effectiveness of the proposed model.

show abstract

“…The paper [44] proposes a novel spatio-temporal model for monitoring the condition of a wind turbine. The k nearest neighbours method was used to construct graphically structured data.…”

Section: Introductionmentioning

confidence: 99%

Stochastic operation model for readiness assessment of small wind turbines based on Markov theory

Zalewska‐Lesiak,

Małachowski,

Szkutnik‐Rogoż

2023

IET Renewable Power Gen

View full text Add to dashboard Cite

Wind energy is now one of the most important alternative sources of renewable energy. The development of wind power generation is a prerequisite to meet environmental and sustainable development requirements while achieving energy security goals. The condition for the effectiveness of the proper operation of a wind power plant is its ability to perform the intended functions, which can be expressed by the coefficient of technical readiness. This article presents an original analysis of the readiness of small wind turbines. As part of the research, an original stochastic operation model was developed based on the application of the Markov process theory. Based on the collected data, a 3‐state phase space of the studied process was identified. Various analyses were carried out to calculate the operating coefficients of wind power plants. The ergodic (boundary) probabilities were calculated using three methods. The conditional probabilities were then determined based on the initial distribution. The resulting readiness factors, not exceeding 30%, indicate the low operational efficiency of the small wind turbines in question. Furthermore, an analysis of the sensitivity of the performance factors was carried out for the analysed objects.

show abstract

Condition monitoring of wind turbines with the implementation of spatio-temporal graph neural network

Cited by 42 publications

References 39 publications

AutoML e transferência de aprendizado por reforço entre problemas de otimização combinatória

AutoML e transferência de aprendizado por reforço entre problemas de otimização combinatória

Multi-Order-Content-Based Adaptive Graph Attention Network for Graph Node Classification

Stochastic operation model for readiness assessment of small wind turbines based on Markov theory

Contact Info

Product

Resources

About