Feature Selection Methods in the Framework of mRMR

Wang, Xiujuan; Tao, Yuanrui; Zheng, Kan

doi:10.1109/imccc.2018.00307

Cited by 6 publications

(1 citation statement)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In other words, we can use Eq. 5 to maximize the differences between univariate relevance and redundancy to determine the added t-th feature (Wang et al, 2018).…”

Section: Modeling Theories 21 Mrmr Feature Selectionmentioning

confidence: 99%

AGC regulation capability prediction and optimization of coal-fired thermal power plants

Jin,

Hao,

Zhang

et al. 2023

Front. Energy Res.

View full text Add to dashboard Cite

The improvement of the AGC regulation capability of thermal power plants is very important for the secure and stable operation of the power grid, especially in the situation of large-scale renewable energy access to the power grid. In this study, the prediction and optimization for the AGC regulation capability of thermal power plants is proposed. Firstly, considering parameters related to the AGC regulation of the thermal power plant, the max-relevance and min-redundancy (mRMR) is used to extract features from historical sequences of the parameters. Next, a model with multi-long short-term neural networks (mLSTM) is constructed to predict the AGC regulation capability; that is, the obtained feature set is considered as the inputs of the first LSTM sub-model to predict future values of the main steam pressure and main steam temperature, which are then utilized as the inputs of the second LSTM sub-model to predict the actual power generation during AGC regulation operation. Then, the AGC regulation index is calculated according to the “management rules of grid-connected operation of power plant in Northern China” and “management rules of auxiliary service of the grid-connected power plant in Northern China” (i.e., “two rules”), and it is then considered as the objective function to be maximized by optimizing the coal feed rate, air supply rate, and feedwater flow rate. Finally, the actual AGC regulation process of a 300 MW coal-fired power plant is used as an application, and the results show that the proposed method can effectively predict and improve the regulation capability when the AGC instruction is received from the power grid.

show abstract

“…In other words, we can use Eq. 5 to maximize the differences between univariate relevance and redundancy to determine the added t-th feature (Wang et al, 2018).…”

Section: Modeling Theories 21 Mrmr Feature Selectionmentioning

confidence: 99%