Elektrikli araçlarda menzil artışına yönelik yapay zekâ tabanlı sürdürülebilir ulaşım modelinin geliştirilmesi, optimizasyonu ve analizi
Küçük Resim Yok
Tarih
2025
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
İnönü Üniversitesi
Erişim Hakkı
info:eu-repo/semantics/openAccess
Özet
Tez çalışmasında, elektrikli araçlarda menzil uzatmak için gerçek zamanlı büyük veriler kullanarak elektrikli aracın kritik parametreleri, yol parametreleri ve coğrafi parametreleri de dahil ederek yeni bir enerji tüketim tahmin modeli geliştirilmiştir. Bu model ile elektrikli araç dönüşümü yapılacak olan rotalarda, elektrikli aracın enerji tüketim tahmini, kullanılacak olan aracın motor tipi ve güç seçimi, rotaya göre batarya kapasitesi seçimi yapılmıştır. Çalışma boyunca büyük veriler kullanılarak enerji tüketim tahmin modelleri elde edilmiştir. Bu modeller oluşturulurken aracın ve rotanın kritik parametreleri arasında yer alan hız, ağırlık, eğim, dış ortam sıcaklığı, elektriksel ve mekaniksel arıza, ivmelenme ile faydalı frenleme gücü kullanılmıştır. Tez çalışması altı adımda ele alınmıştır. İlk adımda literatür taraması yapılmış ve ikinci adımda ise Trambüs'lerin kara kutusunda 1 milyar satır veri toplanmıştır. Üçüncü adımda, BAP doktora tez projesi kapsamında yeni bir kara kutu alınıp araca yerleştirilerek veriler alınmış, eski ve yeni veriler arasında incelemeler yapılmıştır. Dördüncü adımda, Lineer Regresyon yöntemi ile MTECM enerji tüketim tahmin modeli elde edilmiştir. Beşinci adımda GPR kullanılarak GMTECM elde edilmiştir. Son adımda ise SHO-EBECM enerji tüketim tahmin modeli SeaHorse optimizasyon tekniği ile oluşturulmuştur. Modeller, Malatya'da halihazırda otobüs işletmesi yapılan en yoğun rotalara uygulanmış ve işletmenin içten yanmalı araçlar yerine elektrikli araç dönüşümü yapıldığı taktirde rotalar için olası enerji tüketim değerleri hesaplanmıştır. Geliştirilen tüm tahmin modelleri, literatürde yer alan diğer tüketim modelleri ile hata oranı açısından karşılaştırmalı olarak olarak değerlendirilmiş sunulmuştur. Anahtar Kelimeler: Büyük veri, Elektrikli araçlar, Regresyon, SeaHorse, Optimizasyon.
In thesis study, a new energy consumption prediction model was developed by including critical parameters of electric vehicles, road parameters and geographical parameters using real-time big data to extend the range of electric vehicles. With this model, the energy consumption prediction of electric vehicles, the engine type and power selection of the vehicle to be used, and the battery capacity selection according to the route were made in the routes where electric vehicle conversion will be made. Energy consumption prediction models were obtained using big data throughout the study. While creating these models, speed, weight, slope, outside temperature, electrical and mechanical failure, acceleration and recuperation, which are among the critical parameters of the vehicle and route, were used. The thesis study was handled in six steps. In first step, a literature review was conducted and in second step, 1 billion lines of data were collected in black box of Trambuses. In third step, a new black box was purchased within the scope of the BAP project and placed in the vehicle, data was collected, and examinations were made between old and new data. In fourth step, the MTECM energy consumption prediction model was obtained with the Linear Regression method. In fifth step, GMTECM was obtained using GPR. In last step, SHO-EBECM energy consumption estimation model was created with the SeaHorse optimization technique. The models were applied to the busiest routes in Malatya and possible energy consumption values for the routes were calculated if the company converted from fossil vehicles to electric vehicles. All developed prediction models are evaluated and presented comparatively with other consumption models in the literature in terms of error rate. Keywords: Big-Data, Electric vehicles, Regression, SeaHorse, Optimization
In thesis study, a new energy consumption prediction model was developed by including critical parameters of electric vehicles, road parameters and geographical parameters using real-time big data to extend the range of electric vehicles. With this model, the energy consumption prediction of electric vehicles, the engine type and power selection of the vehicle to be used, and the battery capacity selection according to the route were made in the routes where electric vehicle conversion will be made. Energy consumption prediction models were obtained using big data throughout the study. While creating these models, speed, weight, slope, outside temperature, electrical and mechanical failure, acceleration and recuperation, which are among the critical parameters of the vehicle and route, were used. The thesis study was handled in six steps. In first step, a literature review was conducted and in second step, 1 billion lines of data were collected in black box of Trambuses. In third step, a new black box was purchased within the scope of the BAP project and placed in the vehicle, data was collected, and examinations were made between old and new data. In fourth step, the MTECM energy consumption prediction model was obtained with the Linear Regression method. In fifth step, GMTECM was obtained using GPR. In last step, SHO-EBECM energy consumption estimation model was created with the SeaHorse optimization technique. The models were applied to the busiest routes in Malatya and possible energy consumption values for the routes were calculated if the company converted from fossil vehicles to electric vehicles. All developed prediction models are evaluated and presented comparatively with other consumption models in the literature in terms of error rate. Keywords: Big-Data, Electric vehicles, Regression, SeaHorse, Optimization
Açıklama
Anahtar Kelimeler
Elektrik ve Elektronik Mühendisliği, Electrical and Electronics Engineering
