ORIGINAL PAPER
Ball mill load identification method based on IRF-Net with multi-signal time-frequency images
1
School of Mechanical and Electrical Engineering, Jiangxi University of Science and Technology
2
Jiangxi Mechanical and Electrical Engineering Technology Research Center of Mining and Metallurgy
Submission date: 2024-07-24
Final revision date: 2024-09-11
Acceptance date: 2025-02-14
Publication date: 2025-03-19
Corresponding author
Gaipin Cai
School of Mechanical and Electrical Engineering, Jiangxi University of Science and Technology
School of Mechanical and Electrical Engineering, Jiangxi University of Science and Technology
Gospodarka Surowcami Mineralnymi – Mineral Resources Management 2025;41(1):219-237
KEYWORDS
TOPICS
Optimization of resources and reserves managementPreparation and processing of mineral raw materialsRational use of mineral raw materials
ABSTRACT
Accurately identifying the load status of the ball mill during the grinding process is conducive to improving the overall production efficiency and ensuring the safe operation of the entire grinding process. In this study, ball mill loads were classified into nine categories based on charge volume ratio (CVR) and material-to-ball volume ratio (MBVR). Different sensors are utilized to collect cylinder vibration and acoustic signals in the grinding process, respectively, and the raw data are converted into time-frequency images by continuous wavelet transform. In this paper, the ResNet18 model is improved from three aspects, namely, depthwise separable convolution (DSC), dropout layer, and Hardswish activation function, and an improved residual fusion network (IRF-Net) based on the merging of two time-frequency image signals is proposed for load recognition. In order to validate the performance of the proposed model, time-frequency images of the acquired data are analyzed, single and multiple signals are used as network inputs, respectively, compared with other classical models, and ablation experiments are performed on the different modules of the improvement. The results show that the improved residual fusion network achieves the best results in recognition with an accuracy of 98.33%, demonstrating good load recognition. The IRF-Net-based multi-signal time-frequency diagram identification method can be utilized to make a sound judgment on the load status of the mill.
ACKNOWLEDGEMENTS
This research was funded by National Natural Science Foundation of China (52364025).
CONFLICT OF INTEREST
The Authors have no conflict of interest to declare.
METADATA IN OTHER LANGUAGES:
Polish
Metoda identyfikacji obciążenia młyna kulowego oparta na sieci IRF-Net z wielosygnałowymi obrazami czasowo-częstotliwościowymi
obraz czasowo-częstotliwościowy, sieci rezydualne, splot separowalny głębokościowo, sygnały młyna
Dokładne określenie stanu obciążenia młyna kulowego podczas procesu mielenia sprzyja poprawie ogólnej wydajności produkcji i zapewnia bezpieczną pracę całego procesu mielenia. W tym badaniu obciążenia młyna kulowego zostały sklasyfikowane do dziewięciu kategorii na podstawie stosunku objętości wsadu (CVR) i stosunku objętości materiału do kuli (MBVR). Różne czujniki są wykorzystywane do zbierania drgań cylindra i sygnałów akustycznych w procesie mielenia, odpowiednio, a surowe dane są konwertowane na obrazy czasowo-częstotliwościowe za pomocą ciągłej transformacji falkowej. W tym artykule model ResNet18 został ulepszony pod trzema względami, a mianowicie: poprzez zastosowanie splotu separowalnego głębokościowo (DSC), warstwy dropout i funkcji aktywacji Hardswisha, a ulepszona sieć fuzji resztkowej (IRF-Net) oparta na połączeniu dwóch sygnałów obrazu czasowo-częstotliwościowego jest proponowana do rozpoznawania obciążenia. Aby zweryfikować wydajność proponowanego modelu, analizowane są obrazy czasowo-częstotliwościowe pozyskanych danych, pojedyncze i wielokrotne sygnały są używane jako wejścia sieciowe, odpowiednio, w porównaniu z innymi klasycznymi modelami, a eksperymenty ablacji są przeprowadzane na różnych modułach ulepszenia. Wyniki pokazują, że ulepszona sieć fuzji resztkowej osiąga najlepsze wyniki w rozpoznawaniu z dokładnością 98,33%, co świadczy o dobrym rozpoznawaniu obciążenia. Metodę identyfikacji wielosygnałowego diagramu czasowo-częstotliwościowego opartą na IRF-Net można wykorzystać do rzetelnej oceny stanu obciążenia młyna.
REFERENCES (20)
1.
Ak et al. 2022 – Ak, A., Topuz, V. and Midi, I. 2022. Motor imagery EEG signal classification using image processing technique over GoogLeNet deep learning algorithm for controlling the robot manipulator. Biomedical Signal Processing and Control 72(6), DOI: 10.1016/j.bspc.2021.103295.
2.
Ali et al. 2019 – Ali, M.Z., Shabbir, M.N.S.K., Liang, X., Zhang, Y. and Hu, T. 2019. Machine learning-based fault diagnosis for Single- and multi-faults in induction motors using measured stator currents and vibration signals. IEEE Transactions on Industry Applications 55(3), pp. 2378–2391, DOI: 10.1109/TIA.2019.2895797.
3.
Alzubaidi et al. 2021 – Alzubaidi, L., Zhang, J., Humaidi A.J., Ayad Al‑Dujaili, A., Duan, Y., Al‑Shamma, O., Santamaría, J., Fadhel, M.A., Al‑Amidie, M. and aith Farhan, L. 2021. Review of deep learning: concepts, CNN architectures, challenges, applications, future directions. Journal of big Data 8, DOI: 10.1186/s40537-021-00444-8.
4.
Bortnowski, P., Gładysiewicz, L., Król, R. and Ozdoba, M. 2021. Energy Efficiency Analysis of Copper Ore Ball Mill Drive Systems. Energies 14(6), DOI: 10.3390/en14061786.
5.
Cai et al. 2019 – Cai, W.H., Liu, A., Bing, D. and Luo, C. 2019. Load state identification method for ball mills based on improved EWT, multiscale fuzzy entropy and AEPSO_PNN classification. Processes 7(10), DOI: 10.3390/pr7100725.
6.
Gupta, V.K. 2020. Energy absorption and specific breakage rate of particles under different operating conditions in dry ball milling. Powder technology 361, pp. 827–835, DOI: 10.1016/j.powtec.2019.11.033.
7.
He et al. 2019 – He, K.M., Zhang, X.Y., Ren, S.Q. and Sun, J. 2016. Deep residual learning for image recognition. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp. 770–778, DOI: 10.1109/CVPR.2016.90.
8.
Hoang, D.T. and Kang, H.J. 2020. A motor current signal-based bearing fault diagnosis using deep learning and information fusion. IEEE Transactions on Instrumentation and Measurement 69(6), pp. 3325–3333, DOI: 10.1109/TIM.2019.2933119.
9.
Huang et al. 2020 – Huang, P., Sang, G., Miao, Q., Ding, Y. and Jia, M. 2020. Soft measurement of ball mill load based on multi-classifier ensemble modelling and multi-sensor fusion with improved evidence combination. Measurement Science and Technology 32(1), DOI: 015105.10.1088/1361-6501/aba885.
10.
Jin et al. 2024 – Jin, X., Jiang, J., Li, Y. and Wang, Z. 2024. Improved ShuffleNetV2 for Action Recognition in BPPV Treatment. Biomedical Signal Processing and Control 88, DOI: 10.1016/j.bspc.2023.105601.
11.
Kong et al. 2023 – Kong, Y., Wang, X., Zhou, J. and Qin, L. 2023. A Mill Load Identification Method Based on Deep Neural Network. China Automation Congress (CAC). IEEE, pp. 6747–6752, DOI: 10.1109/CAC59555.2023.10452121.
12.
Krizhevsky et al. 2012 – Krizhevsky, A., Sutskever, I. and Hinton, G.E. 2012. ImageNet classification with deep convolutional neural networks. Advances in Neural Information Processing Systems 25(2), pp. 84–90, DOI: 10.1145/3065386.
13.
Liu et al. 2022 – Liu, D., Wang, L., Du, Y., Cong, M. and Li, Y. 2022. 3-D prostate MR and TRUS images detection and segmentation for puncture biopsy. IEEE Transactions on Instrumentation and Measuremen 71, DOI: 10.1109/TIM.2022.3192292.
14.
Shao et al. 2019 – Shao, S., McAleer, S., Yan, R. and Baldi, P. 2019. Highly-Accurate Machine Fault Diagnosis Using Deep Transfer Learning. IEEE Transactions on Industrial Informatics 15(4), pp. 2446–2455, DOI: 10.1109/TII.2018.2864759.
15.
Shi et al. 2020 – Shi, Y., Deng, A., Deng, M., Zhu, J., Liu, Y. and Cheng, Q. 2020. Enhanced lightweight multiscale convolutional neural network for rolling bearing fault diagnosis. IEEE Access 8, pp. 217723–217734, DOI: 10.1109/ACCESS.2020.3041735.
16.
Wang et al. 2019 – Wang, J., Mo, Z., Zhang, H. and Miao, Q. 2019. A deep learning method for bearing fault diagnosis based on time-frequency image. IEEE Access 7, pp. 42373–42383, DOI: 10.1109/ACCESS.2019.2907131.
17.
Wang et al. 2021 – Wang, X., Sun, K., Zhang, H., Xiong, W. and Yang, C. 2021. Mill load identification method for ball milling process based on grinding signal. IFAC-PapersOnLine 54(21), pp. 7–12, DOI: 10.1016/j.ifacol.2021.12.002.
18.
Xu et al. 2022 – Xu, H., Wang, T., Zou, W.J., Zhao, J.J., Tao, L. and Zhang, Z.J. 2022. Ball mill load status identification method based on the convolutional neural network, optimized support vector machine model, and intelligent grinding media. Chinese Journal of Engineering 44(11), pp. 1821–1831, DOI: 10.13374/j.issn2095-9389.2022.03.06.001.
19.
Yang, L. and Yang, H. 2024. Load identification method of ball mill based on the CEEMDAN-wavelet threshold-PMMFE. Gospodarka Surowcami Mineralnymi – Mineral Resources Management 40(2), pp. 163–180, DOI: 10.24425/gsm.2024.150823.
20.
Zhang et al. 2021 – Zhang, K., Wang, J., Shi, H., Xiaochen Zhang, X. and Tang, Y. 2021. A fault diagnosis method based on improved convolutional neural network for bearings under variable working conditions. Measurement 182, DOI: 10.1016/j.measurement.2021.109749.
| eISSN: | 2299-2324 |
| ISSN: | 0860-0953 |
We process personal data collected when visiting the website. The function of obtaining information about users and their behavior is carried out by voluntarily entered information in forms and saving cookies in end devices. Data, including cookies, are used to provide services, improve the user experience and to analyze the traffic in accordance with the Privacy policy. Data are also collected and processed by Google Analytics tool (more).
You can change cookies settings in your browser. Restricted use of cookies in the browser configuration may affect some functionalities of the website.
You can change cookies settings in your browser. Restricted use of cookies in the browser configuration may affect some functionalities of the website.
