Research on heat exhanger fault diagnosis based on HTRI CFD analysis and machine learning technology
Başlık çevirisi mevcut değil.
- Tez No: 795124
- Danışmanlar: Belirtilmemiş.
- Tez Türü: Yüksek Lisans
- Konular: Nükleer Mühendislik, Nuclear Engineering
- Anahtar Kelimeler: Belirtilmemiş.
- Yıl: 2020
- Dil: İngilizce
- Üniversite: Tsinghua University
- Enstitü: Yurtdışı Enstitü
- Ana Bilim Dalı: Belirtilmemiş.
- Bilim Dalı: Belirtilmemiş.
- Sayfa Sayısı: 49
Özet
Özet yok.
Özet (Çeviri)
This study investigates the error warning analysis of heat exchanger data with the machine learning method, which is a recent subject. The heat exchangers, which vary widely according to the purpose of use in the production facilities where high energy productions are made, undertake heating or cooling functions in the system in which they are located. Many different types can be used in many different areas. In the study, using the shell and tube model, these are generally used in areas where the area is expected to be large and heat exchange is high, the water that needs to be heated on the tube side and the steam that gives the evaporation heat to the water in the tube is imitated and transferred to the computer environment. The heat exchanger geometry, which is used to fully imitate the reality, is tried to reach the real data by using CFD method in computer environment. The size of the geometry made it difficult to get results, and a great effort and time was spent to match the data in the computer environment with the field data. This study attempted to imitate one-on-one imitation, as it will further assist us in obtaining erroneous data. Although the geometry is complex, the HTRI program was used to facilitate CFD analysis. No matter how simple the HTRI program analysis simplifies, the methods used to describe the event in the real are incomplete, and the geometry has been simplified due to the large number of tubes in the geometry and huge geometry structure. Morever, steam, which is the condensate on the shell side, increased the analysis time in the computer environment. On the other hand, the model of the heat exchanger, which tries to achieve the results in the CFD environment with the new trend of the world called machine learning, the early warning model has been established for the prediction of a long life and potential accident situation in which the heat exchanger is included. Since machine learning is a very large subject, the methods and programs to be applied here are quite high. These are: linear algebra, Octave / Matlab program, neural networks, support vector machines, supervised learning, unsupervised learning, dimensionality reduction,principal component analysis and anomaly detection which is our main topic for applying on this case. Various platforms and program knowledge were needed to apply these topics, which are the theory parts of machine learning, in computer environment. These platforms are: Anaconda, Jupyter Notebook, Python and SQLite. Thanks to all these elements and their Abstract III assistants, the data had to be analyzed and the result had to be reached. The data was divided into two as test data and training data to perform machine learning, test data was used to test the trained data, and the trained data was used to train the model. In this work, the aim of this study is to increase the safety of the new generation power generation facilities, to detect the faults in these facilities that need to work continuously, by decreasing the risk of accidents by using a computer methodology, to help predict the fault detection before the accident, to take the facilities' automatic operation a step further and to provide extending the life of the equipment by following the data.
Benzer Tezler
- Building sensor-based real-time predictive maintenance system by utilizing artificial intelligent techniques
Yapay akıllı teknikleri kullanarak sensör tabanlı gerçek zaman tahminli bakım sistemi kurulması
RAGHAD MOHAMMED KHORSHEED
Doktora
İngilizce
2021
Endüstri ve Endüstri Mühendisliğiİstanbul Teknik ÜniversitesiEndüstri Mühendisliği Ana Bilim Dalı
DR. ÖĞR. ÜYESİ ÖMER FARUK BEYCA
- Faz değişimli süreçlerde eksergo ekolojik analizin sayısal olarak incelenmesi
Numerical study of exergo environmental analysis at phase change processes
BURAK TEVFİK DOĞAN
Yüksek Lisans
Türkçe
2011
Makine MühendisliğiYıldız Teknik ÜniversitesiMakine Mühendisliği Ana Bilim Dalı
PROF. DR. OLCAY KINCAY
- Borulu tip ısı değiştiricilerinde laminar akışta nümerik olarak ısı geçişi karakteristiklerinin araştırılması
A research on heat transfer characteristics of hairpin heat exchangers in laminar flow numerically
ONUR DALLI
Yüksek Lisans
Türkçe
2019
Makine MühendisliğiYıldız Teknik ÜniversitesiElektronik ve Haberleşme Mühendisliği Ana Bilim Dalı
DOÇ. DR. AHMET SELİM DALKILIÇ
- Fuel oil ile çalışan endüstriyel tip skoç kazanların doğalgaza dönüştürülmesinin bilgisayar simulasyonu
Computer simulation of converting industrial scotch type boilers working with fuel oil into natural gas
ARİF LEVENT KUZAY
- Çamaşır kurutma makinalarındaki havların su ile temizleme tasarımları
Water-powered designs for cleaning the lint in tumbledryers
YUNUS EKREM TEKEŞ
Yüksek Lisans
Türkçe
2019
Makine Mühendisliğiİstanbul Teknik ÜniversitesiMakine Ana Bilim Dalı
PROF. DR. CEVAT ERDEM İMRAK