王龍，現為北京科技大學副教授，香港政府博士獎學金（Hong Kong PhD Fellowship）獲得者，國際電氣和電子工程師協會（IEEE）會員，IEEE Power and Energy Society （IEEE PES）會員，香港城市大學系統工程博士，英國倫敦大學學院（University College London）計算機科學碩士。主要研究方向為數據發掘、機器學習、計算機視覺和計算智能方法在新能源、電力市場和軌道交通領域的應用。研究成果主要發表在IEEE Transactions on Smart Grid, IEEE Transactions on Industrial Informatics 和 IEEE Transactions on Industrial Electronics等SCI期刊，現為SCI期刊Journal of Intelligent Manufacturing, Electronics Letters和IEEE Transactions on Cybernetics審稿人。
個人主頁：http://longwang.me/
Research Experience
Integration of Machine Learning and Computational Intelligence 2016-Present
The goal of this project is to develop a regression model guided swarm intelligence algorithm. We are working on integrating the Gaussian process model into a swarm intelligence algorithm and thus the Gaussian process model is utilized for estimating the fitness function values. This algorithm has been applied to track the maximum power point of PV systems.
GPU-based Parallel Jaya Algorithm 2016-Present
This project aims to develop a parallel Jaya algorithm implemented on GPU. Compared with the conventional Jaya algorithm, the parallel Jaya algorithm is also free of any algorithm-specific parameters and the three main procedures, solution update, fitness value computation, and the best/worst solution selection are all computed in parallel on GPU. We have applied this algorithm to estimate parameters of a li-ion battery model.
Object Detection Using UAVs 2015-Present
Objection detection algorithms are highly desired in emerging applications of UAVs for remote inspection tasks. We are working on designing an improved cascading classifier for objection detection based on UAV-taken images. In order to reduce the number of features utilized, decision trees and support vector machines are utilized as substitutions of boosting algorithms. This algorithm has been applied for detecting surface cracks on wind turbine blades.
Anomaly Detection of Complex Systems 2014-Present
This project aims at developing data-driven anomaly detection approaches for complex systems. We are working on developing deep learning based frameworks to detection anomalies of complex systems. In these frameworks, deep learning algorithms, such as deep autoencoders and dropout deep neural networks, are employed to model complex systems while statistical control charts are utilized to monitor the abnormal statuses. We have applied these frameworks for wind turbine condition monitoring and fault diagnosis.
Short-term Electricity Price Forecasting 2014-2015
Developed an extended stacked denoising autoencoders model, which incorporates both the stochastic neighbor embedding and the random sample consensus algorithms. This model has outperformed classical data-driven models and an industrial method in forecasting electricity prices of five hubs in the USA.
Teaching Experience
Postgraduate Level:
SEEM 6015 Supply Chain Management, Semester A 2015/16, Class Size: 85
Undergraduate Level:
SEEM 4025 Quality Systems & Management, Semester B 2015/16, Class Size: 30
SEEM 3040 Engineering Database \& Systems, Semester A 2016/17, Class Size: 18
Presentations and Invited Talks
"Data-driven Wind Turbine Condition Monitoring,'' 2016 East Lake International Forum for Outstanding Overseas Young Scholars, December, 2016, Wuhan, China
"Data Mining and its Application to Wind Energy,'' China Longyuan Power Group Corporation Ltd., November, 2016, Beijing, China
"Wind Turbine Gearbox Failure Monitoring Based on SCADA Data Analysis,'' 2016 IEEE Power and Energy Society General Meeting, July 2016, Boston, USA
"Wind Turbine Gearbox Failure Monitoring Based on SCADA Data Analysis,'' Seminar Series, Department of Systems Engineering and Engineering Management, City University of Hong Kong, August, 2016, Hong Kong
"Data-driven Wind Turbine Generation Performance Monitoring,'' China Longyuan Power Group Corporation Ltd., August, 2015, Beijing, China
Scholat.com/sira
