Model ayrıştırma temelli oran kontrol yapısı
Model seperation based ratio control structure
- Tez No: 637543
- Danışmanlar: PROF. DR. MÜJDE GÜZELKAYA
- Tez Türü: Yüksek Lisans
- Konular: Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve Kontrol, Computer Engineering and Computer Science and Control
- Anahtar Kelimeler: Belirtilmemiş.
- Yıl: 2020
- Dil: Türkçe
- Üniversite: İstanbul Teknik Üniversitesi
- Enstitü: Fen Bilimleri Enstitüsü
- Ana Bilim Dalı: Kontrol ve Otomasyon Mühendisliği Ana Bilim Dalı
- Bilim Dalı: Kontrol ve Otomasyon Mühendisliği Bilim Dalı
- Sayfa Sayısı: 89
Özet
Bu çalışmada süreç kontrol uygulamalarında karşılaşılan oran kontrol problemi için yeni ve etkin bir yöntem önerilmiştir. . Oran kontrol problemi, iki giriş ve iki çıkış içeren bir kontrol problemi olup, karıştırma işlemleri, yakıt-hava oranının ayarlanması, soğutma düzeneklerinde sıvı buhar oranının ayarlanması gibi uygulamalarda karşılaşılır. İlk örneklerinde basit vanalar yardımı ile çözülen oran kontrol problemleri, günümüzde gıda ve kimya endüstrisinin dramatik bir biçimde büyümesi, bununla birlikte artan talep ve rekabet sebebiyle, özellikle süreç kontrolunda üzerinde yoğun olarak çalışılan bir problem olarak görülmüştür. Süreç kontrol uygulamalarında sıklıkla olarak kullanılan oran kontrol yapıları da yıllar içerisinde geliştirilerek daha etkin yöntemler ortaya çıkartılmıştır. Oran kontrolünde temel amaç, iki süreç değişkeninin birbirleri ile olan oranını sabit tutmaktır. Bu problemde önemli olan aradaki oranın sadece sürekli hal durumunda değil, referans değişimleri ve bozucu etkimeleri durumunda da sabit kalmasını sağlamaktır. Bu yöntemler süreç kontrol uygulamalarında sıklıkla kullanılan matematiksel model olan birinci mertebeden ölü zamanlı modeller için uygun kontrolörlerin tasarlanması ve iki sistem arasındaki oranın referans değişimi ve bozucu etkisi durumunda bu oranın sabit kalmasını sağlamaya yönelik mekanizmalardan oluşur. Bu mekanizmaların ve tasarlanan kontrolörlerin iyileştirilmesi özellikle son yüzyılda büyük önem kazanan verim artışı üzerinde olumlu bir etkiye sahip olacaktır. Çalışmada öncelikle oran kontrol problemi açıklanmış, isterler belirtilmiştir. Ardından bu zamana kadar önerilen oran kontrol yapıları anlatılmış ve incelenmiştir. İncelenen yapıların zayıf noktaları ve üstün oldukları durumlar gösterilmiştir. Ardından, daha önceden yüksek mertebeli sistemlerin ve kaskad sistemlerin kontrolünde kullanılan“Model Ayrıştırma Temelli Kontrolör Tasarımı”yönteminin iç yapısına“Sanal Çevrim Bozucuları”eklenmiş ve elde edilen yeni kontrol yöntemi kullanılarak yeni bir oran kontrol yapısı önerilmiştir. Önerilen yeni oran kontrol yapısı ile sistem bozucuları karşısında istenen oranı koruma başarımı artırılmıştır. Ayrıca bu yapının, diğer yöntemlerin aksine, sadece birinci veya ikinci mertebeden sistem modelleri için değil, üç ve daha yüksek mertebeden sistem modelleri için de uygulanabilir olması, önerilen yöntemin üstünlüğünü göstermektedir. En son bölümde, önerilen yöntem, başka oran kontrol yöntemleriyle karşılaştırılarak, farklı şartlar altındaki üstünlükleri açıklanmıştır.
Özet (Çeviri)
In this study, a new and effective method is proposed for the rate control problem, which is frequently encountered in process control applications. The ratio control problem is a control problem that includes two inlets and two outlets, and it is encountered in applications such as mixing processes, adjusting the fuel-air ratio, and adjusting the liquid vapor ratio in the cooling devices. In this systems the system models are not coupled together. In the first examples, the rate control problems that were solved with the help of simple valves were seen as a problem that has been studied intensively especially in process control due to the dramatic growth of the food and chemical industry today. The first solutions for ratio control problems are consists of a free, or“wild”master flow for first system, that is not controlled for disturbance effects, and a second system that controlled by a controller which has a set point and calculated ratio between system responses. The calculation of the system output ratio directly, brings some difficulties and for application such as zero division and very small denominator value that calculates the ratio extremely high. To overcome this problem, the ratio calculation is made not directly after the measurent of process variables. The second system output is multipled by the desired ratio“a”and drives the controller. These older ratio control systems are simple and very applicable for especially flow control problems. But the wild system has no disturbance rejection and used like only a reference for second system. For more efficency not only to keep the ratio but also to keep both systems in a desired reference, the“wild”system should be controlled. With these two controlled system, the system outputs are controlled to reach the reference and reject disturbance effects. This control maintains the tracking for reference signal for both systems and disturbance rejection so that the desired ratio“a”is preserved during the steady-state. However, due to the increasing demand and competition, ratio control structures, which are used extensively in process control applications, have been developed over the years and more effective methods have been revealed. The main purpose in ratio control applications is to keep the ratio of the two process variables constant to each other and maintain this ratio“a”when reference signal is changed or a disturbance effect occurs to system. The important thing in this problem is to keep the ratio“a”constant not only in the steady state but also in case of reference changes and disruptive effects. These methods mainly consist of mechanisms for designing suitable controllers for first-order plus dead-time models, which are the mathematical models frequently used in especially process control applications, and to ensure that the ratio between the two systems remains constant in case of reference change and disturbance effect. Improvement of these mechanisms and designed controllers will have a positive effect on the increase in efficiency, which has gained great importance in the last century. To overcome the problem of keeping the ratio“a”, stable during the reference change and disturbance effect on one of the systems, pushed the researchers to create effective structures between these two systems. This researches begins with Ratio Station thought. The ratio station is actually a multiplier that creates the reference for second system, that is why the ratio“a”is kept only in steady-state. This solution is not effective when one of the controlled systems affected by an input or output disturbance. Therefore, a dynamic solution should be used to achieve to keep the ratio in transient response, both reference changes and disturbance rejection state. Dynamic solutions are grouped as“blend station”and“dynamic blend station”. These blend stations are create signals with reference signal and output signal of first system and use this signal for second systems reference signal to make sure the both system output signal track the desired reference signal. Blend station solutions are designed with parameters that select the second system reference as mixture of serial and parallel ratio control structure. As an example, if operator select the mixture parameter as 0, system behaves as parallel ratio control structure, on the contrary, the system behaves as serial control structure, while transient response. In the study, firstly, the ratio control problem is explained and the demands are specified for these kinds of ratio control applications. Then, the ratio control structures that are proposed up to this time are explained and examined with their individual block diagrams. The weak points of the examined structures are shown in which cases they are advantageous and in what cases they are disadvantageous and compared with each other. Subsequently, a new ratio control structure was proposed along with the gain values obtained by improving the Model separation-based controller design method that previously used in the control of high-order systems and serial cascade control systems by improving the parameter calculation for this method. The model separation-based controller design basically a design method for higher order systems and used for serial cascade control systems. The main advantage of this method is that the ease of tuning controller with only one parameter, that is closed loop time constant value. The designed controller is not a high order controller, but has n number of simple controllers as P, PI or PID serially, where n represents the total number of control loops in control system. The examination of model separation-based controller design structure comes with a valuable calculated result that includes model nonlinearity and disturbance effects. Also, the other outcome is this method is called virtual loops that leads a way to design basic controllers as P, PI and PID number of n, exchange of higher order complex controllers. These virtual loops make the controlled system like a serial cascade control system. All these virtual loops are treated as a closed loop-controlled system and have controllers to ensure no steady-state error for step responses and effective disturbance rejection performances. The disturbance effected to these virtual loops are called virtual disturbance. Virtual disturbance is not the real system disturbance that caused by physical reasons. Virtual disturbances can be defined only signals for virtual loops, cannot be added to system output directly. The superiority of the proposed method is the dynamism of the desired ratio of protection against the system disturbances of the proposed control structure and the fact that it is applicable not only for the first or second order systems but also for three and higher order systems and also systems that includes deadtime dynamics unlike other methods. The ease of calculation for controller parameters with respect to only one variable is also another advantage for the model separation-based ratio control structure. The other main advantage is that there is no need to make a model reduction to design the controller that is why there are fewer modelling errors, nonlinearities and assumptions with the system model for controller design and real system. In the last section of this study, a simulation comparison for the proposed method with two other ratio control structure methods that have solutions for better ratio control in disturbance effect than classical ratio contol methods. The comparison is made for four different situations as ideal situation that includes perfect modelling and linearity, no modelling error and assumptions and/or reduction, after that modelling error with the first system model as %10 which means all the system parameters gain, pole location and dead-time value estimated %10 lower than actual system , modelling error with second system as %10 , modelling error for both systems as % 10 and finally for higher order systems with controller designed for system model that obtained by model reduction for compared two methods. In last scenario, the proposed method needs no model reduction operation, but the other two compared method has no solution for higher order system models, that is why the controllers for other two methods are designed for models which are derived from step response with an identification method. For a fair comparison, a ratio error index is defined to calculate the achieved success for compared systems. The index is defined for both ratio amplitude and time that the desired ratio is not satisfied even lower difference from desired ratio value. It means that the lower value of this performance index means better performance for compared ratio control structures. The performances for different situations are compared and explained both graphically and numerically. Finally, some of the points that has to be considered while applying proposed ratio control method is shown. Therefore different approach is suggested to reduce the control signal peaks for proposed method, and the outputs of this suggestion are compared and examined by means of same performance index and the results are discussed for the proposed method.
Benzer Tezler
- Karakteristik oran atama yöntemi ile kaskad kontrolör tasarımı
Cascade controller design using characteristic ratio assignment method
AHMET CAN ERDEM
Yüksek Lisans
Türkçe
2021
Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve Kontrolİstanbul Teknik ÜniversitesiKontrol ve Otomasyon Mühendisliği Ana Bilim Dalı
PROF. DR. MÜJDE GÜZELKAYA
- Characterization and modeling of negative-biastemperature instability in 40 NM CMOS technologythrough long short-term memory (LSTM) networks
Uzun kısa-süreli bellek ağlarıyla (LSTM)40 NM CMOS teknolojisinde negatif-kutuplamasıcaklık kararsızlığının karakterizasyonu ve modellenmesi
FİKRET BAŞAR GENCER
Yüksek Lisans
İngilizce
2023
Bilim ve Teknolojiİstanbul Teknik ÜniversitesiElektronik ve Haberleşme Mühendisliği Ana Bilim Dalı
DOÇ. DR. MUSTAFA BERKE YELTEN
- Equimolar binary compounds: A first-principles based machine learning study
Eşmolar ikili alaşımlar: AB-initio tabanlı makina öğrenme çalışması
CEM ORAN
Yüksek Lisans
İngilizce
2021
Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve Kontrolİstanbul Teknik ÜniversitesiHesaplamalı Bilimler ve Mühendislik Ana Bilim Dalı
PROF. DR. ADEM TEKİN
- Solving image procesing problems by using nonstandart regularization
Standart olmayan düzgünleştirme kullanarak görüntü işleme problemlerinin çözümü
TOLGA ACAR
Yüksek Lisans
İngilizce
1994
Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve Kontrolİstanbul Teknik ÜniversitesiDOÇ.DR. MUHİTTİN GÖKMEN
- A numerical investigation into the effect of the vertical tapered winglets on fully-submerged hydofoils
Dikey kanatçıkların tamamen batmış sualtı kanatlarına etkilerinin sayısal yöntemle incelenmesi
ARAS ÇETİNKAYA
Yüksek Lisans
İngilizce
2017
Gemi Mühendisliğiİstanbul Teknik ÜniversitesiGemi İnşaatı ve Gemi Makineleri Mühendisliği Ana Bilim Dalı
DOÇ. DR. UĞUR ORAL ÜNAL