Power Plant Classification from Remote Imaging with Deep Learning

The development of deep learning methodologies for the analysis of thermal power plant load losses requires a combination of real plant data and data derived from fundamental physics-based process models. For this purpose, a robust integrated power plant thermofluid process model of a complete +600MW coal-fired power plant was developed within the Flownex Simulation Environment. It consists of standard and compound components, combined with specially developed scripts to ensure complete energy balance, specifically on the two-phase tank components. This enables simulation of the complete plant operation to determine power output as a function of any given set of internal and external operational variables, boundary conditions and component states. The model was validated against real plant design and acceptance test data. In order to demonstrate the ability of the model it was used to evaluate the plant performance related to three specific load loss inducing scenarios. The results show that a combination of mechanical faults, process anomalies and operational phenomena can be analysed. This provides the basis for generating model-based performance data that can be combined with real plant data to facilitate the development of deep learning analytics tools for load loss fault diagnosis and root cause analysis, as well as fault propagation and load loss forecasting.

Download Full-text

A Procedure Performing System using Deep Learning Algorithm for Nuclear Power Plant

Proceedings of the 30th European Safety and Reliability Conference and 15th Probabilistic Safety Assessment and Management Conference ◽

10.3850/978-981-14-8593-0_4124-cd ◽

2020 ◽

Author(s):

Jeeyea Ahn ◽

Seung Jun Lee

Keyword(s):

Deep Learning ◽

Power Plant ◽

Nuclear Power Plant ◽

Nuclear Power ◽

Learning Algorithm ◽

Deep Learning Algorithm

Download Full-text

Multi-Parameter Prediction for Steam Turbine Based on Real-Time Data Using Deep Learning Approaches

10.1115/gt2021-60049 ◽

2021 ◽

Author(s):

Lei Sun ◽

Tianyuan Liu ◽

Yonghui Xie ◽

Xinlei Xia

Keyword(s):

Neural Network ◽

Deep Learning ◽

Power Plant ◽

Steam Turbine ◽

Real Time ◽

Learning Approaches ◽

Time Data ◽

Time Parameters ◽

Process Prediction ◽

Parameter Prediction

Abstract Accurate and real-time parameters forecasting is of great importance to the turbine control and predictive maintenance which can help the improvement of power system. In this study, deep-learning models including recurrent neural network (RNN) and convolutional neural network (CNN) for multi-parameter prediction are proposed, and are applied to predict real-time parameters of steam turbine based on data from a power plant. Firstly, the prediction results of RNN and CNN models are compared by the overall performance. The two models show good performance on forecasting of six state parameters while RNN performs better. Moreover, the detailed performance on a certain day show that the relative error of two models are both less than 2%. Finally, the influence of model designs including loss function, training size and input time-steps on the performance of RNN model are also explored. The effects of the above parameters on the prediction performance, training and prediction time of the models are studied. The results can provide a reference for model deployment in the power plant. It is convinced that the proposed method has a high potential for dynamic process prediction in actual industrial scenarios through the above research.

Download Full-text

Nuclear power plant sensor signal reconstruction based on deep learning methods

Annals of Nuclear Energy ◽

10.1016/j.anucene.2021.108765 ◽

2022 ◽

Vol 167 ◽

pp. 108765

Author(s):

Zixiao Yang ◽

Peng Xu ◽

Biao Zhang ◽

Chuanlong Xu ◽

Liming Zhang ◽

...

Keyword(s):

Deep Learning ◽

Power Plant ◽

Nuclear Power Plant ◽

Nuclear Power ◽

Signal Reconstruction ◽

Sensor Signal ◽

Learning Methods

Download Full-text

Nuclear Power Plant Thermocouple Sensor Fault Detection and Classification using Deep Learning and Generalized Likelihood Ratio Test

IEEE Transactions on Nuclear Science ◽

10.1109/tns.2017.2697919 ◽

2017 ◽

pp. 1-1 ◽

Cited By ~ 3

Author(s):

Shyamapada Mandal ◽

B. Santhi ◽

S. Sridhar ◽

K. Vinolia ◽

P. Swaminathan

Keyword(s):

Deep Learning ◽

Power Plant ◽

Nuclear Power Plant ◽

Fault Detection ◽

Likelihood Ratio Test ◽

Nuclear Power ◽

Generalized Likelihood Ratio Test ◽

Ratio Test ◽

Sensor Fault ◽

Sensor Fault Detection

Download Full-text

Bi-Objective Dispatch of Multi-Energy Virtual Power Plant: Deep-Learning-Based Prediction and Particle Swarm Optimization

Applied Sciences ◽

10.3390/app9020292 ◽

2019 ◽

Vol 9 (2) ◽

pp. 292 ◽

Cited By ~ 3

Author(s):

Jiahui Zhang ◽

Zhiyu Xu ◽

Weisheng Xu ◽

Feiyu Zhu ◽

Xiaoyu Lyu ◽

...

Keyword(s):

Particle Swarm Optimization ◽

Deep Learning ◽

Power Plant ◽

Short Term Memory ◽

Pareto Frontier ◽

Particle Swarm ◽

Practical Case ◽

Virtual Power ◽

Swarm Optimization ◽

Virtual Power Plant

This paper addresses the coordinative operation problem of multi-energy virtual power plant (ME-VPP) in the context of energy internet. A bi-objective dispatch model is established to optimize the performance of ME-VPP in terms of economic cost (EC) and power quality (PQ). Various realistic factors are considered, which include environmental governance, transmission ratings, output limits, etc. Long short-term memory (LSTM), a deep learning method, is applied to the promotion of the accuracy of wind prediction. An improved multi-objective particle swarm optimization (MOPSO) is utilized as the solving algorithm. A practical case study is performed on Hongfeng Eco-town in Southwestern China. Simulation results of three scenarios verify the advantages of bi-objective optimization over solely saving EC and enhancing PQ. The Pareto frontier also provides a visible and flexible way for decision-making of ME-VPP operator. Two strategies, “improvisational” and “foresighted”, are compared by testing on the Institute of Electrical and Electronic Engineers (IEEE) 118-bus benchmark system. It is revealed that “foresighted” strategy, which incorporates LSTM prediction and bi-objective optimization over a 5-h receding horizon, takes 10 Pareto dominances in 24 h.

Download Full-text