Sharif Digital Repository

Integrity of closed-loop systems with integral action control is characterized by using a series of successively more difficult to obtain conditions: Integral Stabilizable (IS), Integral Controllable (IC), Integral Controllable with Integrity (ICI), and Decentralized Integral Controllability (DIC). The IS condition requires that for a Linear Time-Invariant (LTI) finitedimensional (FD) plant, G(s), one finds a stabilizing decentralized controller with integral action. For G(s) to be IC, in addition to being IS, one should be able to reduce the gain of
all control loops by the same factor from a finite value to (but not including) zero without introducing instabilities. ICI requires that... 

This project focuses on the Tennessee East-man challange problem(TEP) and the impact of setpoint changes and controller tunning on plant performance and profitability. TE has four major units. it is an unstable system due to the reaction within the reactor and a recycle streamcausing snow ball effect. As a result the control of this system is not easy. The study has been conducted along two main directions. First, selecting the optimal values for the setpoints and applying them in the simulation. secondly, try to find better parameter for controller in simulation. Simulation is based on working of Ricker. In TEP like the other multivariable systems, it is ameanable to use a centralized... 

Intraction in multi input multi output (MIMO) system makes control of these systems difficult.There are two ways to control such systems; Centralized and Decentralized control. One way in which a centralized controller is designed is using a precompensator. With application of the pre-compensator, the MIMO sytem is decoupled into several SISO systems and one can control these seperatly. Almost all previews methods for the design of pre-compensator assume that the model of the plant containsno uncertainties.In this thesis a new method is proposed which takes into account model uncertaintes in the design of the pre-compensator.The form of the uncertainties considred an interval-parametric  

Decentralized control is a typical method for the design of industrial MIMO control systems. However, in most multivariable systems, the existence of large interactions limit the achievable performance of such controllers. In the Nyquist Array design method proposed by Rosenbrock, the system interactions can be reduced using a precompensator, such that a decentralized controller may be used with acceptable performance. A well designed pre-compensator will effectively decouple the system and the multivariable systems may be considered as a set of SISO systems. It is therefore obvious that the key element in the Nyquist Array design method is the design of the pre-compensator. There are two... 

Multivariable control systems as used in the industry usually contain integral action in order to improve tracking of the output with respect to step reference changes. These systems are usually subject to input saturation due to actuator or sensor nonlinearities. The aim of this work is find the conditions for the integrity concept of such multivariable system, e.g. IS, IC, ICI and DIC. First we develop the conditions on the controller gains necessary stabilize the systems and then show that these conditions can changes when one examines integrity conditions  

Design of fault tolerant controllers with a reliable and robust control structure that can accommodate sensor and actuator failure is a challenging problem. In the process control literature, the term “integrity” is used to characterize these properties for decentralized control systems. It refers to the ability to retain system stability when one or some of the control loops get in or out of service or when loop gains are detuned.Conditions pertaining to the integrity of the closed-loop system are described by a series of successively more difficult to attain criteria whose evaluation is only based on the process steady state gains. These are Integral stabilizibility (IS), Integral... 

In continous chemical process, disturbances usually propagate along the process. This project describes a data-driven method for identifying the direction of propagation of disturbances using historical process data. The concept is the application of Transfer Entropy, a method based on the conditional probability density functions that measures directionality of variation. For demonstration purposes the transfer entropy method is used to isolate the root-cause of the disturbance in the Tennassee Eastman benchmark process  

The servomechanisms have wide industrial application such as robotics, Computer Numerical Control (CNC) machines, automated manufacturing and solar tracking. Several methods have already been proposed for servo motor control that the most important of them are Proportional Integral Derivative (PID), H2/H--, intelligent based controllers, and model predictive control. Problem specifications and user requirements, will determine the appropriate approach in each case. The aim of this thesis is construction of a real-life camera stabilizer to be used in the motion picture industry. Effects of base movement on the performance should be decreased by choosing appropriate sensors and control... 

In this thesis, the optimal control structure selection and design of sparse multi-variable control systems is addressed. A fundamental challenge which frequently emerges in engineering, social, and economic sciences, is the optimal selection of a subset of elements, in order to maximally fulfil a design objective. In practice, it is required to solve this underlying selection problem in conjunction with a non-linear or non-convex optimization which is designed to ensure desired performance. The requirement to solve these two problems simultaneously is what makes it inherently difficult; one which has thus far eluded efforts to develop a systematic means of determining its solution. In spite... 

Due to the prominence of low-order controllers in industry, this thesis intends to improve the common design algorithms in this field. In this regard, the main challenge is that this problem is NP-hard due to the non-convex rank constraint which appears in the formulation. The existing algorithms mostly solve this problem by obtaining a convex sub-space for this constraint and using output feedback convex optimization methods.Two of common algorithms are non-iterative. These methods two main steps; first to obtain the full-order Lyapunov matrices related to the controller, and then, to fix the null-space of the matrices which have rank constraint. One of the advantages of the algorithm... 

Eigenvalue analysis is a fundamental tool in control issues, however, in some abnormal matrixes it can lead to an incorrect judgment of the matrix dynamic behavior. Pesudospectra, as an eigenvalue generalization, can provide more extensive information of the system dynamic behavior, whereas, the obtained data from eigenvalue analysis can be incomplete. In applying robust controllers, the concept of uncertainty has been attracted researcher interest in system modelling, nevertheless, the controller implementation may not be fault free or the parameters may vary over time. In this project the aforementioned issue will be studied and additionally, the pseudospectra will be employed to design a... 

This research studies the application of coopeartive control and game theoretic approaches in the field of wind farm optimization. The conventional (greedy) wind farm control strategy seeks to individually maximize the power of each turbine.However, due to the presence of aerodynamic interactions (wake effect) between turbines, this strategy will not lead to maximum overall farm power. Accordingly, this study formulates the wind farm power optimization problem as an identical interest game which is utilized to deal with cooperative control problems in general, and the wind farm control problem in particular. Two distributed model-free learning algorithms are then applied to obtain the... 

Optimization-based wind farms control focusing on fatigue loads minimization represents a promising approach to establish a balanced tradeoff between power production in wind farms and lifetime extension of the wind turbines. This dissertation aims to develop a novel strategy to determine individual wind turbine power set-points in such a way that fatigue loads imposed on the units are minimized. The research considers main oscillation modes of the turbine structure and the determining forces to evaluate the amount of fatigues.Based on detailed modelling of wind farms which includes influences of wake effects on the mean wind speed and turbulence intensity, a novel approach is used to... 

In recent years, autonomous unmanned aerial vehicles (UAVs) have attracted considerable amount of interest because of a wide area of applications and a lot of advantages. UAVs have been extensively used for military operations that include tracking, surveillance, active engagement with weapons and airborne data acquisition. UAVs are also in demand commercially due to their advantages in comparison to manned vehicles. These advantages include lower manufacturing and operating costs, flexibility in configuration depending on customer request and not risking the pilot on demanding missions. Safety and efficiency are crucial factors in drones, so in this thesis, these issues are considered. Wind... 

In this thesis, a new method for the design and Control Performance Assessment (CPA) of the Rrestricted-Structure (R-S) PID controllers family (P, I, PI, PD …) is proposed. The proposed method is capable of a systematic analysis of the tradeoff between the structural complexity of the controller and its performance.During the past two decades, control performance assessment has attracted much at- tention from academic and the industrial researchers. One of the most recent tools is the RS-CPA, proposed by Grimble in which restricted structure controllers such as the PID are assessed against optimal LQG measures. When one is able to impose a specific structure onto the controller, such as a... 

Order reduction is a very important issue in Control Theory. A growing need for order reduction models in different fields such as simulation, identification, and design of control system shows this significance. Actually, a high-order system makes a great deal of complexity in designing hardware of control system, debugging, and implementation. Till nowadays, many repetitive as well as nonrepetitive methods with various criteria have been introduced to find low-order models. In this research, order reduction of linear time invariant system models is analyzed. The selected criterion for measuring error between original system and reduced order system is the norm of H1 because it is not only... 

In this thesis we introduce a new methodology to solve the problem of multi-objective control design of linear output-feedback with a constraint on the order of controller. This problem can be formulated as a set of linear matrix inequalities and an additional rank constraint. This constraint is highly non-convex and makes this optimization problem NP-hard. The proposed methodology results in methods which are multi-step and in each step a convex optimization problem shall be solved. This methodology can consider any combination of different design objectives such as H2 performance, H1 performance, passivity and regional pole assignment. In the development of described methodology, a... 

There are a lot of ways to measure the soil compaction but all of them have some deficiencies like low precision, soil destructive, time consuming and etc. The usual ways to measure the compaction in Iran are so traditonal and have the above problems. To solve these problems, modern ways like intelligent compaction is recomonded to decrease the time and cost of compaction. The observor can also control the compaction of all the soil point to prevent any offenses by making the network between the rollers.In new methods the soil compaction is calculated by measuring the roller vibratory and usage of fourier transform from the drum response . To compact the soil more precisely and more quickly,... 

Monitoring of soil moisture plays an essential role in correct decision making and implementation of any closed loop control system in precision agriculture. On the one hand, continuous using of a lot of moisture sensors and monitoring of soil moisture via direct measurement is restricted by practical prohibitions. On the grounds that activation of a lot of sensors continuously requires communication of a large amount of information and causes high power consumption for a moisture monitoring system. On the other hand, using remote sensing methods such as satellite methods, for moisture monitoring in arbitrary spatial and time resolution for control purposes, is not practically possible. The... 

One type of non-stationary environment studied in reinforcement learning involves environments where only one model from a limited set is valid at each time step. The weighted mixture policy method in these environments uses predictions of future model changes to increase cumulative rewards. This method creates a new policy by combining the weighted optimal policy of the current model with that of the future model, after receiving a prediction and before the model change. Obtaining the optimal policy for models is a time-consuming process requiring a large amount of data. This thesis examines existing reinforcement learning methods in non-stationary environments. It demonstrates that the...