A5025005622- Machine Fault Diagnosis Techniques
- Energy Load and Power Forecasting
- Fault Detection and Control Systems
- Water Quality Monitoring and Analysis
- Power System Reliability and Maintenance
- Engineering Diagnostics and Reliability
- Advanced Chemical Sensor Technologies
- Air Quality Monitoring and Forecasting
- Electric Power System Optimization
Hebei University of Technology
2019-2020
Accurate and reliable wind speed forecasting is crucial for farm planning grid operation security. To improve the accuracy of forecasting, a novel combined model proposed in this article. First, complete ensemble empirical mode decomposition adaptive noise (CEEMDAN) permutation entropy (PE) are employed to decompose original time series into sub-series with obvious complexity different; overcome disadvantage weak generalization ability single deep learning method when facing diversiform...
With the continuous elevation of demand for large-scale wind turbines and operation & maintenance cost an increasing interest has been rapidly generated on CM (Condition Monitoring) system. The main components are focus all as they overall lead to high repair costs equipment downtime. Thus, it is difficult make comprehensive assessment in assessment. In present study, intelligent machine learning algorithms adopted mine SCADA (Supervisory Control Data Acquisition) system data WTs (wind...
With the construction of large-scale wind turbines, how to reduce operation and maintenance costs has become an urgent problem be solved. In this paper, by extracting actual data turbine in Supervisory Control Data Acquisition (SCADA) system, Bidirectional R ecurrent Neural Networks (BRNN) is used establish prediction model. By eliminating abnormal points caused accidental factors through box diagram, fault risk threshold components optimized. Then, based on residual between value measured...
With the increasing complexity of wind turbines, current situation high failure rates and maintenance costs has attracted attention power operators. The research on health status monitoring turbines is great significance to development industry. In this study, a novel method for evaluating proposed. fully considers characteristics turbine with high-dimensional nonlinearity. Firstly, Gaussian kernel density estimation Local Outlier Factor (GLOF) used clean data. Secondly, feature parameters...
In view of the complexity pollution gas data, general prediction model can not fully understand its change law, so a concentration time series method based on complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN), approximate entropy (AE), radial basis functional neural network (RBFNN), and back propagation networks (BPNN) is proposed in this paper. Firstly, simplifies data by using CEEMDAN AE. Then, simplified multi-dimensional simple are trained predicted RBFNN. The...