This paper mainly studies the application of deep learning model fusion in temperature monitoring. The research aims to enhance the safety and reliability of power system operations. Given the limitations of traditional monitoring methods in assessing the temperature of overhead line connectors in complex environments, a new approach is proposed: the CNN-DQN algorithm, which combines Convolutional Neural Networks (CNN) and Deep Q-Networks (DQN). The method extracts spatial features of temperature data through CNN and inputs these features into a Markov decision process (MDP) model using DQN to learn and optimize connector temperature monitoring strategies. Experimental results show that the CNN-DQN algorithm has a success rate of 93% in temperature detection and early warning. Compared with traditional models, it has significantly improved accuracy, stability and robustness. This research provides a new and effective method for power system temperature monitoring.
MartinezJGomez-PauARibaJR, et al.On-line health condition monitoring of power connectors focused on predictive maintenance. IEEE Trans Power Deliv2020; 36(6): 3611–3618.
2.
LiuKTangJKWangZ, et al.The application of a novel infrared temperature measurement system in HVDC converter valve equipment connector overheat failure prevention. Procedia Comput Sci2019; 154: 267–273.
3.
WielandtSUhlemannSFiolleauS, et al.Low-power, flexible sensor arrays with solderless board-to-board connectors for monitoring soil deformation and temperature. Sensors2022; 22(7): 2814.
4.
WangFChenZSongG. Monitoring of multi-bolt connection looseness using entropy-based active sensing and genetic algorithm-based least square support vector machine. Mech Syst Signal Process2020; 136(1): 106507.
5.
SahaRBiswasSSarmahS, et al.A working prototype using DS18B20 temperature sensor and Arduino for health monitoring. SN Comp Sci2021; 2: 1–21.
6.
WangHXuZGeX, et al.A junction temperature monitoring method for IGBT modules based on turn-off voltage with convolutional neural networks. IEEE Trans Power Electron2023; 38(8): 10313–10328.
7.
WangFHoSCMSongG. Monitoring of early looseness of multi-bolt connection: a new entropy-based active sensing method without saturation. Smart Mater Struct2019; 28(10): 10LT01.
8.
WangFSongG. Monitoring of multi-bolt connection looseness using a novel vibro-acoustic method. Nonlinear Dynam2020; 100(1): 243–254.
9.
HaoSDaiRFuQ, et al.A robust and adhesive hydrogel enables interfacial coupling for continuous temperature monitoring. Adv Funct Mater2023; 33(33): 2302840.
10.
HoangMLCarratuMPacielloV, et al.Body temperature—indoor condition monitor and activity recognition by mems accelerometer based on IoT-alert system for people in quarantine due to COVID-19. Sensors2021; 21(7): 2313.
11.
SrujanaBChenchireddyK. Electric vehicle charging and temperature monitoring system. In: 2024 5th International Conference on Electronics and Sustainable Communication Systems (ICESC). IEEE, 2024, pp. 54–57.
12.
AlazabMKhanSKrishnanSSR, et al.A multidirectional LSTM model for predicting the stability of a smart grid. IEEE Access2020; 8: 85454–85463.
13.
YuMXuFHuW, et al.Using long short-term memory (LSTM) and internet of things (IoT) for localized surface temperature forecasting in an urban environment. IEEE Access2021; 9: 137406–137418.
14.
LeesTBuechelMAndersonB, et al.Benchmarking data-driven rainfall–runoff models in Great Britain: a comparison of long short-term memory (LSTM)-based models with four lumped conceptual models. Hydrol Earth Syst Sci2021; 25(10): 5517–5534.
15.
KattenbornTLeitloffJSchieferF, et al.Review on convolutional neural networks (CNN) in vegetation remote sensing. ISPRS J Photogrammetry Remote Sens2021; 173: 24–49.
16.
WuPHeYLiY, et al.Online prediction of cutting temperature using self-adaptive local learning and dynamic CNN. IEEE Trans Ind Inf2022; 18(12): 8629–8640.
17.
ArkinEYadikarNXuX, et al.A survey: object detection methods from CNN to transformer. Multimed Tool Appl2023; 82(14): 21353–21383.
18.
BharatiPPramanikA. Deep learning techniques—R-CNN to mask R-CNN: a survey. In: Computational Intelligence in Pattern Recognition: Proceedings of CIPR 2019. Springer, 2020: 657–668.
19.
WangBDongMRenM, et al.Automatic fault diagnosis of infrared insulator images based on image instance segmentation and temperature analysis. IEEE Trans Instrum Meas2020; 69(8): 5345–5355.
20.
GaoYLianfangTQiliang&D. Detection of overheating defects in composite insulators based on Mask R-CNN. China Electric Power2021; 54(1): 135–141.
21.
LuWLiJLiY, et al.A CNN-LSTM-based model to forecast stock prices. Complexity2020; 2020: 1–10.
22.
YusufSAAlshdadiAAAlassafiMO, et al.Predicting catastrophic temperature changes based on past events via a CNN-LSTM regression mechanism. Neural Comput Appl2021; 33(15): 9775–9790.
23.
LiYHeYZhangM. Prediction of Chinese energy structure based on convolutional neural network‐long short‐term memory (CNN‐LSTM). Energy Sci Eng2020; 8(8): 2680–2689.
24.
WangHNLiuNZhangYY, et al.Deep reinforcement learning: a survey. Front Inform Technol Electron Eng2020; 21(12): 1726–1744.
25.
SunYZhangYGuoD, et al.Intelligent distributed temperature and humidity control mechanism for uniformity and precision in the indoor environment. IEEE Internet Things J2022; 9(19): 19101–19115.
26.
DongSWangPAbbasK. A survey on deep learning and its applications. Computer Science Review2021; 40: 100379.
27.
RechtB. A tour of reinforcement learning: the view from continuous control. Annu Rev Control Robot Auton Syst2019; 2: 253–279.
28.
MajidAYSaaybiSFrancois-LavetV, et al.Deep reinforcement learning versus evolution strategies: a comparative survey. IEEE Transact Neural Networks Learn Syst2023; 9(35): 11939–11957.
29.
LvLZhangSDingD, et al.Path planning via an improved DQN-based learning policy. IEEE Access2019; 7: 67319–67330.
30.
LiYWangZXuW, et al.Modeling and energy dynamic control for a ZEH via hybrid model-based deep reinforcement learning. Energy2023; 277: 127627.
31.
JangBKimMHarerimanaG, et al.Q-learning algorithms: a comprehensive classification and applications. IEEE Access2019; 7: 133653–133667.
