Predicting Feed-forward Disturbance based on Deep Learning Network that Combines Multi-time Scales
-
摘要: 为加强过程自动化控制中前馈控制系统的稳定性与可靠性,降低控制过程中时滞的影响。提出一种基于LSTM深度学习网络组合多时间尺度对扰动信号进行预测的方法,在考虑精度的同时也注重预测误差波动对系统产生的冲击。预测结果显示均方根误差(RMSE)在0.0013 ~ 0.0074之间,满足的工程实际要求。其次使用峭度来评定预测误差序列在各时刻的变化波动。最后保证预测精度满足工程要求的同时,通过组合时间尺度的方法来调整预测误差序列在各阶段冲击波动变化,使其在具体的工程应用中达到更优的控制效率与更低的风险 。Abstract: In order to strengthen the stability and reliability of the feedforward system, the influence of time lag in the control process is reduced.A method for predicting disturbance signals based on the LSTM deep learning network that combines multi-time scales is proposed. Both the accuracy and impact of prediction error fluctuation on the controlsystem are considered.The prediction results show that the root mean squares error (RMSE) is between 0.0013 and 0.0074, which meets actual engineering requirements. The kurtosis is used to estimate the fluctuation of the prediction error sequence at each moment.To ensure that the prediction accuracy meets the engineering requirements, the multi-time scale method is jointly used to adjust the impact fluctuation of the prediction error sequence at various stages, so as to achieve better control efficiency and lower risk in specific engineering applications.
-
Key words:
- process control /
- feed-forward system /
- LSTM deep learning network /
- multi-time scale /
- kurtosis
-
表 1 不同尺度序列组合后预测对比
时间尺度 均方根误差 峭度 1 h 0.0015 43.2767 2 h 0.0021 14.2924 组合后 0.0014 24.0936 1 h 0.0015 14.8045 4 h 0.0034 11.1430 组合后 0.0023 12.9696 1 h 0.0013 18.8916 6 h 0.0037 8.5267 组合后 0.0023 7.6944 1 h 0.0014 14.9727 9 h 0.0036 4.9660 组合后 0.0022 4.9261 1 h 0.0013 15.8130 12 h 0.0045 4.7789 组合后 0.0026 4.7239 1 h 0.0014 25.0294 18 h 0.0074 2.9083 组合后 0.0036 2.9310 1 h 0.0013 8.9162 24 h 0.0060 2.8819 组合后 0.0030 2.9710 -
[1] 俞金寿, 孙自强. 过程自动化及仪表[M]. 3版. 北京: 化学工业出版社, 2015: 170-173YU J S, SUN Z Q. Process automation and instrumentation[M]. 3rd ed. Beijing: Chemical Industry Press, 2015: 170-173 (in Chinese) [2] 王雪. 基于时间序列模型的高水平学科预测研究[J]. 情报杂志, 2019, 38(6): 45-49, 117 doi: 10.3969/j.issn.1002-1965.2019.06.008WANG X. Research on high-level subject prediction based on time series model[J]. Journal of Intelligence, 2019, 38(6): 45-49, 117 (in Chinese) doi: 10.3969/j.issn.1002-1965.2019.06.008 [3] SALAZAR L, NICOLIS O, RUGGERI F, et al. Predicting hourly ozone concentrations using wavelets and ARIMA models[J]. Neural Computing and Applications, 2018, 31(8): 4331-4340 [4] TSAI Y C, CHEN J H, WANG J J. Predict forex trend via convolutional neural networks[J]. Journal of Intelligent Systems, 2018, 29(1): 941-958 doi: 10.1515/jisys-2018-0074 [5] LECUN Y, BENGIO Y, HINTON G. Deep learning[J]. Nature, 2015, 521(7553): 436-444 doi: 10.1038/nature14539 [6] CONTRERAS J, ESPINOLA R, NOGALES F J, et al. ARIMA models to predict next-day electricity prices[J]. IEEE Transactions on Power Systems, 2003, 18(3): 1014-1020 doi: 10.1109/TPWRS.2002.804943 [7] ZHANG G P. Time series forecasting using a hybrid ARIMA and neural network model[J]. Neurocomputing, 2003, 50: 159-175 doi: 10.1016/S0925-2312(01)00702-0 [8] 李妮, 江岳春, 黄珊, 等. 基于累积式自回归动平均传递函数模型的短期负荷预测[J]. 电网技术, 2009, 33(8): 93-97, 103LI N, JIANG Y C, HUANG S, et al. Short-term load forecasting based on ARIMA transfer function model[J]. Power System Technology, 2009, 33(8): 93-97, 103 (in Chinese) [9] 张冬青, 宁宣熙, 刘雪妮. 基于RBF神经网络的非线性时间序列在线预测[J]. 控制理论与应用, 2009, 26(2): 151-155ZHANG D Q, NING X X, LIU X N. On-line prediction of nonlinear time series using RBF neural networks[J]. Control Theory & Applications, 2009, 26(2): 151-155 (in Chinese) [10] LI G, SHI J, ZHOU J Y. Bayesian adaptive combination of short-term wind speed forecasts from neural network models[J]. Renewable Energy, 2011, 36(1): 352-359 doi: 10.1016/j.renene.2010.06.049 [11] 傅荟璇, 赵红. MATLAB神经网络应用设计[M]. 北京: 机械工业出版社, 2010: 90-91FU H X, ZHAO H. MATLAB neural network application design[M]. Beijing: Mechanical Industry Press, 2010: 90-91 (in Chinese) [12] GRAVES A. Long short-term memory[M]//GRAVES A. Supervised Sequence Labelling with Recurrent Neural Networks. Berlin, Heidelberg: Springer, 2012: 37-45 [13] 李泽龙, 杨春节, 刘文辉, 等. 基于LSTM-RNN模型的铁水硅含量预测[J]. 化工学报, 2018, 69(3): 992-997LI Z L, YANG C J, LIU W H, et al. Research on hot metal Si-content prediction based on LSTM-RNN[J]. CIESC Journal, 2018, 69(3): 992-997 (in Chinese) [14] 窦珊, 张广宇, 熊智华. 基于LSTM时间序列重建的生产装置异常检测[J]. 化工学报, 2019, 70(2): 481-486DOU S, ZHANG G Y, XIONG Z H. Anomaly detection of process unit based on LSTM time series reconstruction[J]. CIESC Journal, 2019, 70(2): 481-486 (in Chinese) [15] HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural Computation, 1997, 9(8): 1735-1780 doi: 10.1162/neco.1997.9.8.1735 [16] MA X L, TAO Z M, WANG Y H, et al. Long short-term memory neural network for traffic speed prediction using remote microwave sensor data[J]. Transportation Research Part C: Emerging Technologies, 2015, 54: 187-197 doi: 10.1016/j.trc.2015.03.014 [17] GREFF K, SRIVASTAVA R K, KOUTNIK J, et al. LSTM: a search space odyssey[J]. IEEE Transactions on Neural Networks and Learning Systems, 2017, 28(10): 2222-2232 doi: 10.1109/TNNLS.2016.2582924 [18] GRAVES A, SCHMIDHUBER J. Framewise phoneme classification with bidirectional LSTM and other neural network architectures[J]. Neural Networks, 2005, 18(5-6): 602-610 doi: 10.1016/j.neunet.2005.06.042 [19] AMARI S I. Backpropagation and stochastic gradient descent method[J]. Neurocomputing, 1993, 5(4-5): 185-196 doi: 10.1016/0925-2312(93)90006-O [20] KINGMA D P, BA J L. Adam: a method for stochastic optimization[C]//The 3rd International Conference for Learning Representations. San Diego, 2015 [21] YEUNG S, RUSSAKOVSKY O, JIN N, et al. Every moment counts: dense detailed labeling of actions in complex videos[J]. International Journal of Computer Vision, 2018, 126(2-4): 375-389 doi: 10.1007/s11263-017-1013-y [22] DUCHI J, HAZAN E, SINGER Y. Adaptive subgradient methods for online learning and stochastic optimization[J]. Journal of Machine Learning Research, 2011, 12(7): 257-269 [23] 张键. 机械故障诊断技术[M]. 北京: 机械工业出版社, 2014: 60-62ZHANG J. Mechanical fault diagnosis technology[M]. Beijing: Mechanical Industry Press, 2014: 60-62 (in Chinese) [24] HSU C W, CHANG C C, LIN C J. A practical guide to support vector classification[EB/OL]. (2016-05-19) [2017-03-20]. https://www.csie.ntu.edu.tw/~cjlin/papers/guide/guide.pdf.