Abstract Stochastic resonance can utilize the energy of noise to enhance weak frequency characteristic. This paper proposes an adaptive multi-stable stochastic resonance method assisted by the neural network (NN) and physics supervision (directly numerical simulation of the physical system). Different from traditional adaptive algorithm, the evaluation of the objective function (i.e., fitness function) in iteration process of adaptive algorithm is through a trained neural network instead of the numerical simulation. It will bring a dramatically reduction in computation time. Considering predictive bias from the neural network, a secondary correction procedure is introduced to the reevaluate the top performers and then resort them in iteration process through physics supervision. Though it may increase the computing cost, the accuracy will be enhanced. Two examples are given to illustrate the proposed method. For a classical multi-stable stochastic resonance system, the results show that the proposed method not only amplifies weak signals effectively but also significantly reduces computing time. For the detection of weak signal from outer ring in bearings, by introducing a variable scale coefficient, the proposed method can also give a satisfactory result, and the characteristic frequency of the fault signal can be extracted correctly.
Corresponding Authors:
Ming Xu
E-mail: xuming@cjlu.edu.cn
Cite this article:
Xucan Li(李栩灿), Deming Nie(聂德明), Ming Xu(徐明), and Kai Zhang(张凯) Adaptive multi-stable stochastic resonance assisted by neural network and physical supervision 2025 Chin. Phys. B 34 050203
[1] Benzi R, Sutera A and Vulpiani A 1981 J. Phys. A: Math. Gen. 14 L453 [2] Mantegna R N and Spagnolo B 1996 Phys. Rev. Lett. 76 563 [3] Van den Broeck C, Parrondo J M R and Toral R 1994 Phys. Rev. Lett. 73 3395 [4] Li D X, Xu W, Yue X L and Lei Y M 2012 Nonlinear Dyn. 70 2237 [5] Anishchenko V S, Neiman A B, Moss F and Shimansky-Geier L 1999 Phys. Usp. 42 7 [6] McDonnell M D and Abbott D 2009 PLoS Comput. Biol. 5 e1000348 [7] Palabas T, Torres J J, Perc M and Uzuntarla M 2023 Chaos, Solitons and Fractals 168 113140 [8] He F and Yang Y 2021 Neuroscience 458 213 [9] Förster A, Guderian A, Zeyer K P, Dechert G and Schneider F W 1996 Int. J. Neur. Syst. 7 385 [10] Leonard D S and Reichl L E 1994 Phys. Rev. E 49 1734 [11] Lu Z Q, Chen L Q, Brennan M J, Yang T J, Ding H and Liu Z G 2016 J. Sound Vib. 370 221 [12] Dykman M I, Luchinsky D G, Mannella R, McClintock P V E, Stein N D and Stocks N G 1995 Il Nuovo Cimento D 17 661 [13] Lindner B and Schimansky-Geier L 2000 Phys. Rev. E 61 6103 [14] Liu Y J, Wang F Z, Liu L and Zhu Y M 2019 Eur. Phys. J. B 92 168 [15] Liao Z Q, Ma K J, Shamim Sarker M, Yamahara H, Seki M and Tabata H 2022 Results Phys. 42 105968 [16] Gong X L, Xu P F, Liu D and Zhou B L 2023 Chaos, Solitons and Fractals 172 113534 [17] Mitaim S and Kosko B 1998 Proc. IEEE 86 2152 [18] Liu X L, Liu H G, Yang J H, Litak G, Cheng G and Han S 2017 Mech. Syst. Signal Process. 96 58 [19] Lei Y G, Han D, Lin J and He Z J 2013 Mech. Syst. Signal Process. 38 113 [20] Li J M, Chen X F and He Z J 2013 Mech. Syst. Signal Process. 36 240 [21] Hu N Q, Chen M and Wen X S 2003 Mech. Syst. Signal Process. 17 883 [22] Guo W, Zhou Z M, Chen C and Li X 2017 Microelectron. Reliab. 75 239 [23] Li J M, Zhang J F, Li M and Zhang Y G 2019 Mech. Syst. Signal Process. 114 128 [24] Jin Y F, Wang H T, Xu P F and Xie W X 2023 Probab. Eng. Mech. 72 103418 [25] Jin Y F, Wang H T and Zhang L L 2024 Chin. Phys. B 33 010501 [26] He L F, Liu Q L and Zhang T Q 2022 Chin. Phys. B 31 070503 [27] McNamara B and Wiesenfeld K 1989 Phys. Rev. A 39 4854 [28] Gang H, Qing G R, Gong D C and Weng X D 1991 Phys. Rev. A 44 6414 [29] Dykman M I, Luchinsky D G, Mannella R, McClintock P V E, Stein N D and Stocks N G 1993 J. Stat. Phys. 70 463 [30] Huang D W, Yang J H, Zhang J L and Liu H G 2017 Proc. Inst. Mech. Eng. C 231 3964 [31] Michalewicz Z and Schoenauer M 1996 Evol. Comput. 4 1 [32] Bhoskar M S T, Kulkarni M R O K, Kulkarni M R N K, Patekar M S S L, Kakandikar G M and Nandedkar V M 2015 Mater. Today: Proc. 2 2624 [33] Katoch S, Chauhan S S and Kumar V 2021 Multimed. Tools Appl. 80 8091 [34] Wang J, Zhang Q, Liang N, Zhang Y Z and Xu G H 2010 Journal of Xi’an Jiaotong University 44 32 [35] Yang J H and Zhou D J 2020 Re-scaled Resonance Theory and Application in Fault Diagnosis (Beijing: Science Press) p. 76 [36] Feldman M 2011 Mech. Syst. Signal Process. 25 735
Altmetric calculates a score based on the online attention an article receives. Each coloured thread in the circle represents a different type of online attention. The number in the centre is the Altmetric score. Social media and mainstream news media are the main sources that calculate the score. Reference managers such as Mendeley are also tracked but do not contribute to the score. Older articles often score higher because they have had more time to get noticed. To account for this, Altmetric has included the context data for other articles of a similar age.