تعهد نامه

مجله پژوهش در بهداشت محیط

شماره جاری

بر اساس شماره‌های نشریه

بر اساس نویسندگان

بر اساس موضوعات

نمایه نویسندگان

نمایه کلیدواژه ها

درباره نشریه

اهداف و چشم انداز

اعضای هیات تحریریه

اصول اخلاقی انتشار مقاله

بانک ها و نمایه نامه ها

پیوندهای مفید

پرسش‌های متداول

فرایند پذیرش مقالات

اخبار و اعلانات

Editorial Policy

Advertising Policy

پیش‌بینی غلظت PM2.5 با استفاده از روش ترکیبی تجزیه ی و شبکه عصبی

مقالات آماده انتشار

نوع مقاله : Research Paper

نویسندگان

1 دانش‌آموخته‌ی کارشناسی ارشد محیط‌زیست، دانشکده منابع طبیعی، دانشگاه کردستان،کردستان، ایران.

2 دانشیار گروه محیط‌زیست، دانشکده منابع طبیعی، دانشگاه کردستان، کردستان، ایران.

چکیده

زمینه و هدف: یکی از شاخص های مهم در بحث کیفیت هوا، غلظت ذرات معلق PM2.5 می باشد. بدین منظور در این پژوهش از مدل ترکیبی تبدیل موجک گسسته حداکثر هم پوشانی-تجزیه مد متغیر–شبکه عصبی پس انتشار (MODWT-VMD-BPNN) بر پایه تکنیک تجزیه دو مرحله‌ای برای پیش‌بینی ذرات‌معلق PM2.5 شهر ارومیه استفاده شده است. 

مواد و روش ها: سری داده‌های اصلی ذرات معلق PM2.5 ابتدا توسط مدل تبدیل موجک گسسته حداکثر هم پوشانی به دوسطح جزئیات با فرکانس بالا (d1 و d2) و یک سطح تقریب با فرکانس پایین (a2) تجزیه و در مرحله دوم هر کدام از سطوح جزئیات و سطح تقریب توسط مدل تجزیه مد متغیر به 8 مود متغیر تجزیه شد. سپس هر کدام از مدهای متغیر توسط شبکه عصبی پس انتشار مدلسازی و پیش بینی شدند. 

یافته ها: بر اساس نتایج مدل های ترکیبی (MODWT-BPNN) و (VMD- BPNN) نسبت به مدل تکی شبکه عصبی پس انتشار (BPNN) عملکرد بهتری داشته اند. و در بین مدل های ترکیبی مدل (MODWT-BPNN) به دلیل تحلیل سیگنال هایی دارای تغییرات ناگهانی و ناپیوستگی موضعی به وسیله ی موجک ها عملکرد بهتری نسبت به مدل (VMD-BPNN) دارد. مدل ترکیبی تجزیه دو مرحله ای ((MODWT-VMD-BPNN نسبت به دیگر مدل های تجزیه ای تک مرحله ای و مدل تکی شبکه عصبی پس انتشار با مقادیر معیارهای ارزیابی خطا شامل 2/8582= RMSE=3/8074 MAE و آماره 0/92 = R در مرحله آموزش و 2/1840 = RMSE = 2/7679  MAE و آماره 0/80= R در مرحله آزمون، عملکرد بهتری داشته است.

نتیجه گیری: مدل های تجزیه ای دو مرحله ای با تجزیه سطح جزئیات و سطح تقریب به 8 مود می تواند مشکل آمیختگی مدها را حل کند و پیش بینی میزان غلظت PM2.5 را با دقت بهتری انجام دهد لذا این مدل می تواند برای پیش بینی آلاینده های جوی به کار گرفته شود.

کلیدواژه‌ها

عنوان مقاله [English]

PM2.5 Concentration Forecasting Using Decomposition Method and Artificial Neural Network

نویسندگان [English]

  • Salah Baizidi 1
  • Jamil Amanollahi 2

1 Master Student, Department of Environmental Sciences, Faculty of Natural Resources, University of Kurdistan, Iran.

2 Associate professor at Department of Environmental Sciences, Faculty of Natural Resources, University of Kurdistan, Iran.

چکیده [English]

Background and purpose: One of the key indicators in discussions about air quality is the concentration of PM2.5 particulate matter. This research employs a combined model that utilizes the maximum overlap discrete wavelet transform, variational mode decomposition, and backpropagation neural network (MODWT-VMD-BPNN). This two-stage decomposition technique aims to predict PM2.5 levels in the city of Urmia.

Material and Methods: Data on air quality in Urmia City, including levels of particulate matter (PM10 and PM2.5), carbon dioxide (CO2), sulfur dioxide (SO2), nitrogen dioxide (NO2), and nitrogen monoxide (NO), were obtained from the General Directorate of Environmental Protection for the years 2019 to 2023. Meteorological data were sourced from the General Directorate of Meteorology of West Azerbaijan Province. In the first stage of the analysis, the original PM2.5 data series was decomposed into two high-frequency detail levels (d1 and d2) and one low-frequency approximation level (a2) using the Maximum Overlap Discrete Wavelet Transform (MODWT) model. In the second stage, each of these detail and approximation levels was further decomposed into eight variable modes using the Variable Mode Decomposition (VMD) model. Subsequently, each variable mode was simulated and predicted using a backpropagation neural network (BPNN). To evaluate the accuracy and performance of the proposed model, it was compared with the MODWT-BPNN, VMD-BPNN, and standard BPNN models.

Results: After reviewing the results, the MODWT-VMD-BPNN model achieved R=0.92, RMSE=3.8074, and MAE=2.8582 during training, and R=0.80, RMSE=2.7679, and MAE=2.1840 during testing, demonstrating superior accuracy and performance compared to the other models.

Conclusion: The two-stage decomposition models tackle mode mixing effectively and enhance the extraction and prediction of multiple frequencies in PM2.5 data with greater precision.
 
Open Access Policy: This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. To view a copy of this licence, visit https://creativecommons.org/licenses/by/4.0/

کلیدواژه‌ها [English]

  • Prediction
  • Particulate Matter PM2.5
  • Wavelet Transform
  • Variational Mode Decomposition
  • Back Propagation Neural Network
مراجع
  1. Esmaili, R. and Amini, F. L. 2021. Identification of hot spots PM2.5 in Mashhad air pollution. Journal of Climate Research. 44:63-68. (In Persian).
  2. Dabiri, M. 2007. Environmental pollution (Air, Water, Soil, Sound). Etehat Tehran, pp400. (In Persian).
  3. Mohammadian, M., Sojodi, L. and Etemadinejad, S. 2011. Survey of concentrations of PM2.5 inside and outside of shops in Sar's city center. Journal of Mazandaran University of Medical Sciences, 21: 72-79. (In Persian).
  4. Deng, X., Zhang, F. and Rui, W. 2013. PM2.5 -induced oxidative stress triggers autophagy in human lung epithelial A549 cells. Taxicology inVitro, 27(6): 17662-1770. https://doi.org/10.1016/j.tiv.2013.05.004 PMid:23685237
  5. Mckendry, I. G. 2000. PM10 levels in the lower Fraser valley, British Columbia, Canada: an overview of spatiotemporal variation and meteorological control, Journal of The Air and Waste Management Association, 50(3): 443-452. https://doi.org/10.1080/10473289.2000.10464025 PMid:10734716
  6. Abidin, A. U., Munawaroh, A. L., Rosinta, A., Sulistiyani, A, Ardianta, i. and Iresha, F. M. 2025. Environmental health risks and impacts of PM2.5 exposures on human health in residential areas, Bantul, Yogyakarta, Indonesia. Toxicology Reports: 14: 101949. https://doi.org/10.1016/j.toxrep.2025.101949 PMid:40026480 PMCid:PMC11869533
  7. Wu, Ch-H., Dang, T. T. M., Mutuku, J. K., Lin, L-M., Huang, B-W. and Chang-Chien, G-P. 2025. Evaluation of PM2.5 bound microplastics and plastic additives in several cities in Taiwan: Spatial distribution and human health risk. Science of The Total Environment, 959: 178213.. https://doi.org/10.1016/j.scitotenv.2024.178213
  8. Zeng, Y., Yang, X., Zhang, A., Yuan, X., Zhai, J., Xing, C., Cai, B., Shi, S., Zhang, Y. and Zhang, Y. 2024. Source-specific health effects of internally exposed organics in urban PM2.5 based on human serum albumin adductome analysis. Science of The Total Environment, 955: 176958.. https://doi.org/10.1016/j.scitotenv.2024.176958 PMid:39419214
  9. Madanchi, P., Saeediyan, H. and Damizadeh, T. M. 2024. Evalution of meteorological data and satellite images in identifying dust phenomenon in desert areas (case study: Kerman province). Journal of Geography and Environmental Hazards, 13(4). (In Persian). https://doi.org/10.22067/geoeh.2024.87528.1476.
  10. Najjarian, M., Etemadfard, H. and Golmohammadi, M. 2024. Identification of the dust formation center with Tensor approach (Study Case: Sabzevar County). Journal of Geography and Environmental Hazards 13(2). (In Persian). https://doi.org/10.22067/geoeh.2023.82992.1384.
  11. Negahban, S., Ganjaeian, H., Ghaysarian, S. A. and Ebrahimi, A. 2024. Identifying the centers of dust and analyzing the factors influencing its occurrence based on remote sensing data (Case Study: Southwest Iran). Journal of Geography and Environmental Hazards, 13(4). (In Persian). https://doi.org/10.22067/geoeh.2024.89088.1504.
  12. Zhou, Q., Jiang, H., Wang, J. and Zhou, J. 2014. A hybridmodel for PM2.5 forecasting based on ensemble empirical mode decomposition and a general regression neural nrtwork. Science of the Total Environment, 496: 264-274. https://doi.org/10.1016/j.scitotenv.2014.07.051 PMid:25089688
  13. Liu, H., Tian, H. Q. and Li, Y. F. 2015. Four wind speed multi-step forecasting models using extreme learning machines and signal decomposing algorithms. Energy Convers. Manag. 100, 16-22. https://doi.org/10.1016/j.enconman.2015.04.057
  14. Feng, X., Li, Q., Zhu, Y., Hou, J., Jin, L. and Wang, J. 2015. Artificial neural network forecasting of PM2.5 pollution using air mass trajectory based geographic model and wavelet transformation. Atmos Environ, 107: 118-128. https://doi.org/10.1016/j.atmosenv.2015.02.030
  15. Bai, Y., Li, Y., Wang, X.X., Xie, J.J. and Li, C. 2016. Air pollutants concentrations forecasting using back propagation neural network based on wavelet decomposition with meteorological conditions. Atmospheric Pollution Research. 7, 557-566. https://doi.org/10.1016/j.apr.2016.01.004
  16. Wang, D., Liu, Y., Luo, H., Yue, Ch. and Cheng, Sh. 2017. Day-ahead PM2.5 concentration forecasting using WT-VMD based decomposition method and back propagation neural network improved by differential evolution. Int J Environ Res Public Health, 14(7): 764. https://doi.org/10.3390/ijerph14070764 PMid:28704955 PMCid:PMC5551202
  17. Ghimire, S., Deo, R. C., Casillas-Perez, D., Sharma, E., Salcedo-Sanz, S., Barua, P. D. and Acharya, U. R. 2024. Half-hourly electricity price prediction with a hybrid convolution neural network-random vector functional link deep learning approach. Applied Energy, 374: 123920. https://doi.org/10.1016/j.apenergy.2024.123920
  18. Fang, J., Yan, Z., Lu, X. and Zhao, Z. 2024. An oil production prediction approach based on variational mode decomposition and ensemble learning model. Computers& Geosciences, 193: 105734. https://doi.org/10.1016/j.cageo.2024.105734
  19. Xue, H., Wang, G., Li, X. and Du, F. 2024. Predictive combination model for CH4 separation and CO2 sequestrarion with CO2 injection into coal seams: VWD-STA-BiLSTM-ELM hybrid neural network modeling. Energy, 313: 133744. https://doi.org/10.1016/j.energy.2024.133744
  20. Su, T., Liu, D., Cui, X., Dou, X., Lei, B., Cheng, X., Yuan, M. and Chen, R. 2024. Prediction of DEDI index for meteorological drought with the VMD-CBiLSTM hybrid model. Journal of Hydrology. 641:131805. https://doi.org/10.1016/j.jhydrol.2024.131805
  21. Rahimpour, A., Amanollahi, J. and Tzanis, C. G. 2021. Air quality data series estimation based on machine learning approaches for urban environments. Air Quality, Atmosphere & Health. 14" 191-201. https://doi.org/10.1007/s11869-020-00925-4
  22. Ghasemi, A. and Amanollahi, J. (2019). Integration of ANFIS model and forward selection method for air quality forecasting. Air Quality and Atmospheric Health.12: 59-72. https://doi.org/10.1007/s11869-018-0630-0
  23. Raei, R., Mohammadi, S. and Fendereski, H. 2015. Forecasting stock index with neural network and wavelet transform. Journal of AssetManagement and Financing. 1: 55-74. (In Persian). 20.1001.1.23831170.1394.3.1.4.4
  24. Dragomiretskiy, K. and Zosso, D. 2014. Variational mode decomposition. IEEE T Signsl Proces, 62: 531-544. https://doi.org/10.1109/TSP.2013.2288675
  25. Tahmasebi, M. K., Kahoo, A. R. and Kalateh, A. N. 2017. Ground-roll noise suppression by time-frequency filtering based on variational mode decomposition. Journal of Research on Applied Geophysics. 3. 177-188. (In Persian). https://doi.org/10.1016/j.jappgeo.2018.10.025.
  26. Khalkhali Shandiz, S. and Khezrzadeh, H. 2021. Application of moving vehicle response and variational mode decomposition (VMD) for indirect damage detection in bridges. Modares Civil Engineering Journal. 21 (1), 31-46. (In Persian). 20.1001.1.24766763.1400.21.1.13.6.
  27. Quiroz, R. C., Yarleque, Ch., Posadas, A., Mares, V. and Immerzeel, W. W. 2011. Improving daily rainfall estimation from NDVI using a wavelet transform. Environmental Modelling and Software, 26(2): 201-209. https://doi.org/10.1016/j.envsoft.2010.07.006
  28. Daubechies, I. 1992. Ten lectures on wavelet. Society for Industrial Mathematics. https://doi.org/10.1137/1.9781611970104 PMid:18296155
  29. Karbasi, M and Dindar, S. 2019. Comparison of wavelet-MLP and wavelet-GMDH models in forecasting EC and SAR at Zayandeh-Rood RIVER. Environmental Sciences. 16 (4), 135-152. (In Persian).
مجله پژوهش در بهداشت محیط

مقالات آماده انتشار، پذیرفته شده
انتشار آنلاین از تاریخ 28 اسفند 1403
فایل ها
هم رسانی
ارجاع به این مقاله
آمار