基于NG-RC电机系统的混沌预测

doi:10.19596/j.cnki.1001-246x.8615

摘要/Abstract

摘要：

基于下一代储备池计算实现电机系统混沌预测, 并能够基于已有数据预测未知变量数据。与传统储备池计算相比, 下一代储备池计算所采用数据本身直接连接, 需要的训练数据集更小。并且下一代储备池计算通过高维转换避免了传统储备池网络复杂的参数优化计算, 使得计算速度实现大幅提升。这一研究结果为电机系统混沌预测提供一种研究思路。

关键词: 下一代储备池计算, 永磁同步电机系统, 混沌预测

Abstract:

In this paper, chaos prediction of motor system is realized based on the next generation reservoir computing, and unknown variable data can be predicted based on existing data. Compared with the traditional reservoir computing, the next generation reservoir computing uses direct connection of data itself and requires smaller training data sets. And the next generation reservoir computing avoids the complex parameter optimization calculation of traditional reserve pool network through high-dimensional conversion, which greatly improves the computing speed. This research result provides a new research idea for chaotic prediction of motor systems.

Key words: next generation reservoir computing, permanent-magnet synchronous motor, chaos prediction

曹文鑫, 韦笃取. 基于NG-RC电机系统的混沌预测[J]. 计算物理, 2023, 40(4): 519-526.

Wenxin CAO, Duqu WEI. Chaos Prediction of Motor System Based on NG-RC[J]. Chinese Journal of Computational Physics, 2023, 40(4): 519-526.

图/表 7

图1 储备池模型结构图(a)传统储备池计算RC模型结构；(b) 下一代储备池计算NG-RC模型结构

Fig.1 Reservoir computing model structure (a) a traditional reservoir computing, RC; (b) next generation reservoir computing (nonlinear vector autoregression), NG-RC

图2 利用NG-RC方法对电机系统进行预测

Fig.2 Forecasting PMSM system using the NG-RC

图3 测试阶段状态变量误差同步误差分析

Fig.3 Synchronization error of state variables in test phase

图4 利用NG-RC对PMSM系统id变量进行预测

Fig.4 Inference PMSM variables id using NG-RC

表1 LSTM、PSO-LSTM、NG-RC结果汇总

Table 1 The best results of LSTM、PSO-LSTM、NG-RC

模型	测试数据集大小	RMSE
LSTM	10 000	0.025 4
PSO-LSTM	10 000	0.010 1
NG-RC	400	0.002 1

图5 PMSM变量id的回归图

Fig.5 PMSM return map of id

表2 NG-RC相对于传统RC计算的速度提升

Table 2 Estimated speed up of the NG-RC for the PMSM forecasting task

	M_warmup	M_train	N_nonlin	N_total	N	σ_r	Speed up
NG-RC	2	400	21	28
LC-RC	1 000	1 000		100	100	0.01~0.05	33~163
Le-RC	0	5 000		300	300	0.02	1.5×10³
L-RC	10⁵	6×10⁴		4 000	2 000	0.02	3.2×10⁶

参考文献 28

1	ZHONG G X , WEI D Q , ZHANG B . Remote chaotic synchronous control of inductive load in distributed power generation system[J]. Chinese Journal of Computational Physics, 2019, 36 (6): 719- 725.
2	LIU L H , WEI D Q , ZHANG B . Synchronization control of chaos in complex motor networks with small-world topology based on dynamic relaying[J]. Chinese Journal of Computational Physics, 2018, 35 (6): 750- 756.
3	LU Y D , WEI D Q . Chaotic time series prediction of power system by using optimized time spectrum neural network[J]. Chinese Journal of Computational Physics, 2022, 39 (3): 371- 378.
4	ZHENG L , LIU Z , SHEN J , et al. Very short-term maximum Lyapunov exponent forecasting tool for distributed photovoltaic output[J]. Applied Energy, 2018, 229 (4): 1128- 1139.
5	ZHANG C T , MA Q L , PENG H . Chaotic time series prediction based on information entropy optimized parameters of phase space reconstruction[J]. Acta Phys Sin, 2010, 59 (11): 7623- 7629. DOI
6	LIU C , LIU C , SHANG Y , et al. An adaptive prediction approach based on workload pattern discrimination in the cloud[J]. Journal of Network and Computer Applications, 2017, 80 (6): 35- 44.
7	ZHOU Y H , LIU H Q , QI P , et al. Forecast of oil production in fractured-vuggy reservoir by using recurrent neural networks[J]. Chinese Journal of Computational Physics, 2018, 35 (6): 668- 674.
8	GRIFFITH A , POMERANCE A , GAUTHIER D J . Forecasting chaotic systems with very low connectivity reservoir computers[J]. Chaos, 2019, 29 (12): 123108. DOI
9	LYMBURN T , KHOR A , STEMLER T , et al. Consistency in echo-state networks[J]. Chaos, 2019, 29 (2): 023118. DOI
10	PATHAK J , HUNT B , GIRVAN M , et al. Model-free prediction of large spatiotemporally chaotic systems from data: A reservoir computing approach[J]. Physical Review Letters, 2018, 120 (2): 024102. DOI
11	GAUTHIER D J , BOLLT E , GRIFFITH A , et al. Next generation reservoir computing[J]. Nature Communications, 2021, 12 (1): 1- 8. DOI
12	HART A G , HOOK J L , DAWES J . Echo state networks trained by Tikhonov least squares are L2(μ) approximators of ergodic dynamical systems[J]. Physica D, 2021, 421 (5): 132882.
13	GONON L , ORTEGA J P . Fading memory echo state networks are universal[J]. Neural Networks, 2021, 138 (11): 10- 13.
14	JAEGER H , HAAS H . Harnessing nonlinearity: Predicting chaotic systems and saving energy in wireless communication.[J]. Science, 2004, 304 (5667): 78- 80. DOI
15	MAASS W , NATSCHLAGER T , MARKRAM H . Real-time computing without stable states: A new framework for neural computation based on perturbations[J]. Neural Computation, 2002, 14 (11): 2531- 2560. DOI
16	CHEN H C , WEI D Q . Chaos prediction and synchronization of motor system based on reservoir computing[J]. Vibration and Shock, 2021, 40 (16): 199- 203.
17	SORIANO M C , ORTÍN S , KEUNINCKX L , et al. Delay-based reservoir computing: Noise effects in a combined analog and digital implementation[J]. IEEE transactions on neural networks and learning systems, 2014, 26 (2): 388- 393.
18	JAURIGUE L , ROBERTSON E , WOLTERS J , et al. Reservoir computing with delayed input for fast and easy optimisation[J]. Entropy, 2021, 23 (12): 1560. DOI
19	WU D , YI Y S , ZHANG Y X . A brief review of integrated and passive photonic reservoir computing systems and an approach for achieving extra non-linearity in passive devices[J]. Science China (Information Sciences), 2020, 63 (6): 97- 104.
20	REN K , ZHANG W Y , WANG F , et al. Next-generation reservoir computing based on memristor array[J]. Acta Phys Sin, 2022, 71 (14): 140701. DOI
21	ZHONG G X , WEI D Q , ZHANG B . Chaos prediction and synchronization of motor system based on reservoir computing[J]. Vibration and Shock, 2020, 39 (7): 8- 14.
22	WEI D Q , LUO X S , WANG B H , et al. Robust adaptive dynamic surface control of chaos in permanent magnet synchronous motor[J]. Physics Letters A, 2007, 363 (1-2): 71- 77. DOI
23	WEI D Q , ZHANG B . Controlling chaos in permanent magnet synchronous motor based on finite-time stability theory[J]. Chinese Physics B, 2009, 18 (4): 1399- 1403.
24	YANG Y F , LUO M Z , XING S B , et al. Analysis of chaos in permanent magnet synchronous generator and optimal output feedback H∞ control.[J]. Acta Phys Sin, 2015, 64 (5): 050503.
25	QING X , NIU Y . Hourly day-ahead solar irradiance prediction using weather forecasts by LSTM[J]. Energy, 2018, 148, 461- 468.
26	REN X , LIU S , YU X D , et al. A method for state-of-charge estimation of lithium-ion batteries based on PSO-LSTM[J]. Energy, 2021, 234, 121236.
27	LU Z , HUNT B R , OTT E . Attractor reconstruction by machine learning[J]. Chaos, 2018, 28 (6): 061104.
28	PATHAK J , LU Z , HUNT B R , et al. Using machine learning to replicate chaotic attractors and calculate Lyapunov exponents from data[J]. Chaos, 2017, 27 (12): 121102.

[1]	刘凯歌, 韦笃取. 基于WOA-ESN的电机系统混沌振荡预测[J]. 计算物理, 2022, 39(4): 498-504.
[2]	卢英东, 韦笃取. 基于遗传注意力机制的DLSTM电力系统混沌预测[J]. 计算物理, 2022, 39(3): 371-378.