Asansör sistemlerinin trafik analizi dizaynı ve simülasyonu
Başlık çevirisi mevcut değil.
- Tez No: 55585
- Danışmanlar: PROF.DR. HAMİT ÖZTEPE
- Tez Türü: Doktora
- Konular: Makine Mühendisliği, Mechanical Engineering
- Anahtar Kelimeler: Asansör, Elevator
- Yıl: 1996
- Dil: Türkçe
- Üniversite: İstanbul Teknik Üniversitesi
- Enstitü: Fen Bilimleri Enstitüsü
- Ana Bilim Dalı: Belirtilmemiş.
- Bilim Dalı: Belirtilmemiş.
- Sayfa Sayısı: Belirtilmemiş.
Özet
Neural networks are usually used for pattern recognition and prediction. It is proposed here that they are used for passenger rate prediction in lifts. It is thus necessary to gather enough data to be used for training. The simulation program is executed on a pseudo building with 9 floors excluding the main terminal. Three cars with various contact capacity are used to service the whole building. In this program interfloor distance, contract speed, time for opening and closing the doors, passenger transfer time are chosen by users. The program runs on a 486PC with 8 MB RAM, 256MB hard disk configuration. It is run for different building configurations with maximum 9 floors and 3 cars. This limitation depends on computer's configuration and software limits as it is shown in SO and et.al. They used their simulation program on 25 floor building with four cars and it has taken 35 hours to train up neural network by using the 6000 records. It is shown how simulation can be used to predict lift system behavior, and a large number of simulation runs are carried out, tabulated and plotted. The ratio between the passenger arrival rate and the handling capacity of the lift system is suggested as a measure of system loading. As a result, average waiting time, round trip time, and interval graphics are obtained by using a simulation program. The possible use of control theory in lift systems is outlined in detail. XIX
Özet (Çeviri)
1. Classification: An input pattern is passed to the network and the network produces a representative class as output. 2. Pattern matching : an input pattern is passed to the network and the network produces the corresponding output pattern. 3. Pattern completion: an incomplete pattern is passed to the network and the network produces an output pattern that has the missing portions on the input pattern filled in. 4. Noise removal: a noise-corrupted input pattern is presented to the network and the network removes some or all of the noise and produces a cleaner version of the input pattern as output. 5. Optimization: An input pattern representing the initial values for a specific optimization problem is presented to the network and the network produces a set of variables that represent a solution to the problem. 6. Control: An input pattern represents the current state of a controller and the desired response for the controller and the output is the proper command sequence that will create the desired response. In this context, it is important to point out that, when using the term Lift Traffic Design, the word design pertains to the selection of lift configuration and not to the mechanical or electrical aspects of the lift system. Whenever the term Average Waiting Time (AWT) is mentioned, it always means Passenger Average Waiting Time, which is the average of all the waiting times that the passenger have to wait, until they board the car. Although that is sometimes used synonymously with System Response Time (SRT) there is a difference. The term Round Trip Time (RTT) is the average period of time for a single elevator car trip around a building, usually during up-peak traffic conditions, measured from the time the car doors open at the main terminal, until the car doors reopen at the main terminal, when the car returns to the main terminal, after its trip around the building. The Interval (INT) is defined as the time between the arrivals of cars at the main terminal. The symbol S is the average number of stop made by the lift in one journey, and H is the average highest reversal floor. The term Handling Capacity (HC) when used, relates to the Lift System Handling Capacity, and is defined as the percentage of the building population served in five minutes. This work is an exploration in the possibilities of the various fields in which neural networks can be applied in lift control. Neural networks are usually applied in problems where no set of clear and decisive rules exist, and where the nature of the problem requires intuitive reasoning, like pattern recognition problems. winNeural networks are usually used for pattern recognition and prediction. It is proposed here that they are used for passenger rate prediction in lifts. It is thus necessary to gather enough data to be used for training. The simulation program is executed on a pseudo building with 9 floors excluding the main terminal. Three cars with various contact capacity are used to service the whole building. In this program interfloor distance, contract speed, time for opening and closing the doors, passenger transfer time are chosen by users. The program runs on a 486PC with 8 MB RAM, 256MB hard disk configuration. It is run for different building configurations with maximum 9 floors and 3 cars. This limitation depends on computer's configuration and software limits as it is shown in SO and et.al. They used their simulation program on 25 floor building with four cars and it has taken 35 hours to train up neural network by using the 6000 records. It is shown how simulation can be used to predict lift system behavior, and a large number of simulation runs are carried out, tabulated and plotted. The ratio between the passenger arrival rate and the handling capacity of the lift system is suggested as a measure of system loading. As a result, average waiting time, round trip time, and interval graphics are obtained by using a simulation program. The possible use of control theory in lift systems is outlined in detail. XIX1. Classification: An input pattern is passed to the network and the network produces a representative class as output. 2. Pattern matching : an input pattern is passed to the network and the network produces the corresponding output pattern. 3. Pattern completion: an incomplete pattern is passed to the network and the network produces an output pattern that has the missing portions on the input pattern filled in. 4. Noise removal: a noise-corrupted input pattern is presented to the network and the network removes some or all of the noise and produces a cleaner version of the input pattern as output. 5. Optimization: An input pattern representing the initial values for a specific optimization problem is presented to the network and the network produces a set of variables that represent a solution to the problem. 6. Control: An input pattern represents the current state of a controller and the desired response for the controller and the output is the proper command sequence that will create the desired response. In this context, it is important to point out that, when using the term Lift Traffic Design, the word design pertains to the selection of lift configuration and not to the mechanical or electrical aspects of the lift system. Whenever the term Average Waiting Time (AWT) is mentioned, it always means Passenger Average Waiting Time, which is the average of all the waiting times that the passenger have to wait, until they board the car. Although that is sometimes used synonymously with System Response Time (SRT) there is a difference. The term Round Trip Time (RTT) is the average period of time for a single elevator car trip around a building, usually during up-peak traffic conditions, measured from the time the car doors open at the main terminal, until the car doors reopen at the main terminal, when the car returns to the main terminal, after its trip around the building. The Interval (INT) is defined as the time between the arrivals of cars at the main terminal. The symbol S is the average number of stop made by the lift in one journey, and H is the average highest reversal floor. The term Handling Capacity (HC) when used, relates to the Lift System Handling Capacity, and is defined as the percentage of the building population served in five minutes. This work is an exploration in the possibilities of the various fields in which neural networks can be applied in lift control. Neural networks are usually applied in problems where no set of clear and decisive rules exist, and where the nature of the problem requires intuitive reasoning, like pattern recognition problems. winNeural networks are usually used for pattern recognition and prediction. It is proposed here that they are used for passenger rate prediction in lifts. It is thus necessary to gather enough data to be used for training. The simulation program is executed on a pseudo building with 9 floors excluding the main terminal. Three cars with various contact capacity are used to service the whole building. In this program interfloor distance, contract speed, time for opening and closing the doors, passenger transfer time are chosen by users. The program runs on a 486PC with 8 MB RAM, 256MB hard disk configuration. It is run for different building configurations with maximum 9 floors and 3 cars. This limitation depends on computer's configuration and software limits as it is shown in SO and et.al. They used their simulation program on 25 floor building with four cars and it has taken 35 hours to train up neural network by using the 6000 records. It is shown how simulation can be used to predict lift system behavior, and a large number of simulation runs are carried out, tabulated and plotted. The ratio between the passenger arrival rate and the handling capacity of the lift system is suggested as a measure of system loading. As a result, average waiting time, round trip time, and interval graphics are obtained by using a simulation program. The possible use of control theory in lift systems is outlined in detail. XIX1. Classification: An input pattern is passed to the network and the network produces a representative class as output. 2. Pattern matching : an input pattern is passed to the network and the network produces the corresponding output pattern. 3. Pattern completion: an incomplete pattern is passed to the network and the network produces an output pattern that has the missing portions on the input pattern filled in. 4. Noise removal: a noise-corrupted input pattern is presented to the network and the network removes some or all of the noise and produces a cleaner version of the input pattern as output. 5. Optimization: An input pattern representing the initial values for a specific optimization problem is presented to the network and the network produces a set of variables that represent a solution to the problem. 6. Control: An input pattern represents the current state of a controller and the desired response for the controller and the output is the proper command sequence that will create the desired response. In this context, it is important to point out that, when using the term Lift Traffic Design, the word design pertains to the selection of lift configuration and not to the mechanical or electrical aspects of the lift system. Whenever the term Average Waiting Time (AWT) is mentioned, it always means Passenger Average Waiting Time, which is the average of all the waiting times that the passenger have to wait, until they board the car. Although that is sometimes used synonymously with System Response Time (SRT) there is a difference. The term Round Trip Time (RTT) is the average period of time for a single elevator car trip around a building, usually during up-peak traffic conditions, measured from the time the car doors open at the main terminal, until the car doors reopen at the main terminal, when the car returns to the main terminal, after its trip around the building. The Interval (INT) is defined as the time between the arrivals of cars at the main terminal. The symbol S is the average number of stop made by the lift in one journey, and H is the average highest reversal floor. The term Handling Capacity (HC) when used, relates to the Lift System Handling Capacity, and is defined as the percentage of the building population served in five minutes. This work is an exploration in the possibilities of the various fields in which neural networks can be applied in lift control. Neural networks are usually applied in problems where no set of clear and decisive rules exist, and where the nature of the problem requires intuitive reasoning, like pattern recognition problems. winNeural networks are usually used for pattern recognition and prediction. It is proposed here that they are used for passenger rate prediction in lifts. It is thus necessary to gather enough data to be used for training. The simulation program is executed on a pseudo building with 9 floors excluding the main terminal. Three cars with various contact capacity are used to service the whole building. In this program interfloor distance, contract speed, time for opening and closing the doors, passenger transfer time are chosen by users. The program runs on a 486PC with 8 MB RAM, 256MB hard disk configuration. It is run for different building configurations with maximum 9 floors and 3 cars. This limitation depends on computer's configuration and software limits as it is shown in SO and et.al. They used their simulation program on 25 floor building with four cars and it has taken 35 hours to train up neural network by using the 6000 records. It is shown how simulation can be used to predict lift system behavior, and a large number of simulation runs are carried out, tabulated and plotted. The ratio between the passenger arrival rate and the handling capacity of the lift system is suggested as a measure of system loading. As a result, average waiting time, round trip time, and interval graphics are obtained by using a simulation program. The possible use of control theory in lift systems is outlined in detail. XIX
Benzer Tezler
- Çok kabinli asansör sisteminin PC ile otomasyonu
Başlık çevirisi yok
UĞUR KESEN
Yüksek Lisans
Türkçe
1993
Elektrik ve Elektronik MühendisliğiYıldız Teknik ÜniversitesiElektronik Mühendisliği Ana Bilim Dalı
PROF. DR. SEZGİN ALSAN
- Mekanik ve elektronik kontrollü asansörlerin verim açısından analizi
Productive analysis of elevators controlled mechanically and elektronically
CUMA ZAFER DİLCİ
Yüksek Lisans
Türkçe
1999
Makine MühendisliğiNiğde ÜniversitesiMakine Mühendisliği Ana Bilim Dalı
YRD. DOÇ. DR. MENDERES KALKAT
- Çerçeve sistemlerinin bilgisayar ile boyutlandırılması ve asansör kabinlerine uygulanması
Computer aided frame system design and its application on the elevator car frame
NİHAT GÜR
- Structured development methodologies for real-time systems
Gerçek zaman sistemleri için yapısal geliştirme metodolojileri
ABDULHAFİD S. SHUWEHDİ
Doktora
İngilizce
1991
Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve KontrolOrta Doğu Teknik ÜniversitesiPROF.DR. ZİYA AKTAŞ
- Asansör simülasyonu ve konfor kartının gerçeklenmesi
Evelator simulation and realization of control device
AYHAN AKÇAİR
Yüksek Lisans
Türkçe
1997
Elektrik ve Elektronik MühendisliğiKaradeniz Teknik ÜniversitesiElektrik-Elektronik Mühendisliği Ana Bilim Dalı
DOÇ. DR. RIFAT YAZICI