
تعداد نشریات | 162 |
تعداد شمارهها | 6,623 |
تعداد مقالات | 71,546 |
تعداد مشاهده مقاله | 126,902,911 |
تعداد دریافت فایل اصل مقاله | 99,952,899 |
توسعه شبکه عصبی مصنوعی و الگوریتم ازدحام ذرات برای پیشبینی جریان ورودی به سدها تحت تأثیر سناریوهای اقلیمی | ||
مدیریت آب و آبیاری | ||
دوره 14، شماره 4، اسفند 1403، صفحه 845-862 اصل مقاله (1.64 M) | ||
نوع مقاله: مقاله پژوهشی | ||
شناسه دیجیتال (DOI): 10.22059/jwim.2024.376149.1160 | ||
نویسندگان | ||
مهرنوش هدایتی زاده؛ سعید جمالی* ؛ هومن حاجی کندی؛ سمیه یوسفی | ||
گروه مهندسی عمران، دانشکده فنی و مهندسی، دانشگاه آزاد اسلامی واحد تهران مرکز، تهران، ایران. | ||
چکیده | ||
تغییر اقلیم با ایجاد تغییر در میزان دما و بارش موجب تغییر در آبدهی رودخانهها میشود. از اینرو، شبیهسازی جریان رودخانه بهعنوان پیشنیاز برنامهریزی و مدیریت منابع و مصارف آب اهمیت فراوان دارد. لذا در پژوهش حاضر تأثیر تغییر اقلیم بر میزان دبی رودخانه مهاباد در دورههای زمانی آیندۀ (2026-2045) با استفاده از مدلهای یادگیری ماشین بررسی شد. ابتدا دو سناریوی ورودی که در آن سناریوی اول شامل پارامترهای دما و بارش و سناریوی دوم شامل پارامترهای دما، بارش و دبی یک ماه قبل بود، تدوین شد. در ادامه عملکرد دو مدل ANN و ANN-PSO در تخمین دبی جریان در دوره پایه (1992-2014) مقایسه شد تا بهترین سناریو و بهترین مدل برای پیشبینی جریان در دوره آینده تحت سه سناریو SSP1.26، SSP2.45 و SSP5.85 گزارش ششم تغییر اقلیم (CMIP6) انتخاب شود. نتایج معیارهای ارزیابی خطا نشان داد که مدل ANN-PSO با استفاده از سناریوی دوم و با معیارهای 77/0=NSE، MCM 4/6=RMSE و MCM 4/3=MAE قادر به تخمین مناسب دبی میباشد. نتایج بررسی اثر تغییر اقلیم بر روی هر یک از پارامترهای هواشناسی نشان داد که تغییر اقلیم باعث افزایش دما در حدود 5/0 تا 1 درجه در طول دوره و ایجاد یک الگوی نوسانی در بارش میشود. نتایج بررسی تغییر اقلیم روی دبی نشان داد که تحت سناریوی SSP1.26 تغییرات چندانی اکثر ماهها در دبی رخ نخواهد داد اما در سناریوهای SSP2.45 و SSP5.85 در ماه دسامبر افزایش اندک دبی رخ خواهد داد و در ماه می و آوریل بیشترین کاهش دبی بهترتیب MCM 5/16 و MCM 33/13 خواهد بود. | ||
کلیدواژهها | ||
پیشبینی سری زمانی؛ تخمین دبی رودخانه؛ تغییر اقلیم؛ ANN؛ PSO | ||
عنوان مقاله [English] | ||
Development of artificial neural network and particle swarm algorithm to predict inflow to dams under the influence of climate scenarios | ||
نویسندگان [English] | ||
Mehrnoosh Hedayatizadeh؛ Saeed Jamali؛ hooman hajikandi؛ Somayeh Yousefi | ||
Department of Civil Engineering, Central Tehran Branch, Islamic Azad University, Tehran, Iran. | ||
چکیده [English] | ||
Climate change causes changes in the flow of rivers by causing changes in temperature and precipitation. Therefore, river flow simulation is important as a prerequisite for some environmental and engineering issues. In the current research, the effect of climate change on the Mahabad’s river flow in the future periods (2045-2026) was predicted using machine learning models. First, two input scenarios were compiled, in which the first scenario included temperature and precipitation parameters and the second scenario included temperature, precipitation, and flow parameters one month ago. In the following, the performance of two ANN and ANN-PSO models in estimating the flow rate in the base period (1992-2014) was compared to select the best scenario and the best model for predicting the flow in the future period under the three scenarios SSP1.26, SSP2.45 and SSP5.85 of the CMIP6. The results of the error evaluation criteria showed that the ANN-PSO model makes the best estimation of the river flow using the second scenario and with the criteria (NSE=0.77, RMSE=6.4 MCM, MAE=3.4 MCM for the test data) and it was chosen to predict the flow in the future period (2026-2045). The results of investigating the effect of climate change on each of the meteorological parameters showed that climate change causes an increase in temperature and creates a fluctuating pattern in precipitation. The results of the climate change survey on flow showed that under the SSP1.26 scenario, there will not be much changes in flow in almost months, but in the SSP2.45 and SSP585 scenarios, there will be an increase in the discharge in December, and in May and April, the greatest decrease in discharge will be (16.50 MCM) and (13.33 MCM) respectively. | ||
کلیدواژهها [English] | ||
Artificial neural network, Climate change, particle swarm algorithm, River discharge, Time series prediction | ||
مراجع | ||
| ||
آمار تعداد مشاهده مقاله: 148 تعداد دریافت فایل اصل مقاله: 93 |