Visual Analytics of RNN for Thermal Power Control System Identification

被引：0
|
作者：
Ji L. [1 ,2 ]
Yang Y. [1 ,2 ]
Qiu S. [1 ,2 ]
Wang Y. [1 ,2 ]
Tian B. [3 ,4 ]
机构：
[1] Beijing Key Laboratory of Petroleum Data Mining, China University of Petroleum, Beijing
[2] Department of Computer Science and Technology, China University of Petroleum, Beijing
[3] Beijing Engineering Research Center of Power Station Automation, Beijing
[4] China Energy Zhishen Control Technology Company Limited, Beijing
来源：
Jisuanji Fuzhu Sheji Yu Tuxingxue Xuebao/Journal of Computer-Aided Design and Computer Graphics | 2021年 / 33卷 / 12期
关键词：
Recurrent neural networks; System identification; Thermal power control; Visual analysis;
D O I：
10.3724/SP.J.1089.2021.19268
中图分类号：
学科分类号：
摘要：
Due to the problems such as strong continuity and high complexity of the data generated by the thermal power control process, and the difficulty of establishing a semantic correlation between the model behavior of recurrent neural networks and the actual control process, the model debugging, optimization and semantic analysis cannot be carried out intuitively. Visual analytics in recurrent neural networks modeling for system identification is ap-plied and a visual analysis system called iaRNN is proposed. Firstly, by visualizing the activation value distribu-tion and coverage of hidden units, the combined view for model evaluation is designed to support the evaluation of model performance in many aspects including inside and outside. Then, visual views are designed from the perspectives of temporal evolution, sensitivity analysis, etc, to support the exploration of model response behavior to thermal power control parameters. Finally, based on the symbolic representation of time sequences and clustering analysis, a method for exploring association patterns between strong time-dependent real-valued time series and hidden units is proposed. A case study using real power plant data is conducted to verify the effectiveness of iaRNN in assisting users to understand the working mechanism of the model and diagnose model defects. © 2021, Beijing China Science Journal Publishing Co. Ltd. All right reserved.
引用
收藏
页码：1876 / 1886
页数：10
相关论文
共 26 条
  • [1] Duan Yanjie, Lv Yisheng, Zhang Jie, Et al., Deep learning for control: the state of the art and prospects, Acta Automatica Sinica, 42, 5, pp. 643-654, (2016)
  • [2] Liu Jizhen, Wang Qinghua, Fang Fang, Et al., Data-driven-based application architecture and technologies of smart power generation, Proceedings of the CSEE, 39, 12, pp. 3578-3587, (2019)
  • [3] Ji Lianen, Kong Yumeng, Wang Yanlin, Et al., Visual analytics of time series data of thermal power control for system identification, Journal of Computer-Aided Design & Computer Graphics, 31, 10, pp. 1677-1686, (2019)
  • [4] Ming Y, Cao S Z, Zhang R X, Et al., Understanding hidden memories of recurrent neural networks, Proceedings of the IEEE Conference on Visual Analytics Science and Technology, pp. 13-24, (2017)
  • [5] Strobelt H, Gehrmann S, Pfister H, Et al., LSTMVis: a tool for visual analysis of hidden state dynamics in recurrent neural networks, IEEE Transactions on Visualization and Computer Graphics, 24, 1, pp. 667-676, (2018)
  • [6] Shen Q M, Wu Y H, Jiang Y Z, Et al., Visual interpretation of recurrent neural network on multi-dimensional time-series forecast, Proceedings of the IEEE Pacific Visualization Symposium, pp. 61-70, (2020)
  • [7] Pascanu R, Mikolov T, Bengio Y., On the difficulty of training recurrent neural networks, Proceedings of the 30th International Conference on Machine Learning, pp. 1310-1318, (2013)
  • [8] Hochreiter S, Schmidhuber J., Long short-term memory, Neural Computation, 9, 8, pp. 1735-1780, (1997)
  • [9] Cho K, van Merrienboer B, Gulcehre C, Et al., Learning phrase representations using RNN encoder-decoder for statistical machine translation, Proceedings of the Conference on Empirical Methods in Natural Language Processing, pp. 1724-1734, (2014)
  • [10] Shi X J, Chen Z R, Wang H, Et al., Convolutional LSTM network: a machine learning approach for precipitation nowcasting, Proceedings of the 28th Neural Information Processing Systems, pp. 802-810, (2015)
123