Catalin Dimon

Control and Optimization for Steel Plant Preheating Installations

Abstract The paper presents the results of the research performed by the authors on control systems and optimization for the operating process of the heating installations from the blast furnace at a steel plant. This system was developed on two relevant levels interconnected in a hierarchical control structure. The acquisition and control level

Polynomial RST Control for Blood Pressure Regulation

This paper proposes polynomial robust RST control for the regulation of the blood pressure in post cardiac surgery patients. Results using Reinforcement Learning is also presented. Clinical applications require simple and effective controllers. This studys focus is to show that polynomial robust RST control is suitable for the blood pressure regulation in post cardiac surgery patients. A realistic and detailed model of a hypertensive patient is used, this model offers a drug response (here we use the model of Slate et al. (1980)), a model for an internal reflex of the body (the baroreceptor reflex) and disturbances. We report simulation results obtained by using polynomial robust RST control, because it ensures the tolerance of the system to the structure disturbances of the model, as well as to the inherent approximations. A result obtained with Reinforcement Learning is also presented.

Blood Pressure Regulation -- Robust Control

This paper proposes a numerical control strategy for the automatic control of the blood pressure in post cardiac surgery patients. The model used presents itself as a system with distributed parameters (SDP). Motivated by the need for simple and effective controllers for clinical applications, this studys main claim is showing that the proposed solution, a robust polynomial R-S-T controller solution, is suitable for the blood pressure regulation in post cardiac surgery patients. A detailed and realistic model of a hypertensive patient is used, this model consists of the drug response model of Slate et al. (1980), for an internal reflex of the body (for which we will propose an improvement) and disturbances. A result obtained with the classical PID control is also presented. To support our main claim, we report simulation results obtained by running a numeric R-S-T controller. A robustness analysis is made. Hence, the results in simulation ensure the tolerance of the system to the structure of the model.

Modeling, Control and Robustness Analysis for a Road Sector

Abstract When modeling a real process the effects of nonlinearities must be taken into account, especially when designing a control algorithm. Such is the case of a road traffic sector, which will be considered from a macroscopic point of view. An input-output model is proposed, based on experimental data, and then a control algorithm is used to obtain imposed performances for the closed loop system. The robustness of control is taken into account, by assessing the modulus margin. Based on this indicator the robustness of the control law is analyzed in order to find what nonlinearities it can cope with. Following, we try to prove that the system can be controlled using the same algorithm, when different types of perturbations affect the system.

Multimodel Control of Diesel Engines

In this article it is proposed and designed a modern control configuration of the type multicontroler-multimodel (MM) that pilots the nonlinear combustion process of the Diesel engine, needed to adjust the pressure in the intake manifold and the airflow circulating through the compressor. The MM simulator developed by the authors allows the implementation of control systems represented by pairs (Mi, Ci) with the Mi candidate closest to the current operating point of the process and the paired controller Ri, for controlling the key parameters of the combustion process. The proposed configuration is built with robust controllers and thus it is able to ensure superior performance, tolerance to nonlinearities and parametric and structural perturbations in the system.

Fuzzy modeling and control for a road section

When dealing with road traffic modeling, one problem is given by the difficulty of modeling the dynamics of a vehicle/group of vehicles from the time it enters a road section until it reaches the end. Besides, road traffic, seen as a complex dynamic process, is difficult to be characterized by fixed values. Therefore, this paper proposes a solution using fuzzy logic control, which by analyzing the occupancy of a section, establishes a number of linguistic features. For modeling the behavior of the road section the macroscopic approach is used.

Decentralized modeling of road traffic

Road traffic is considered as a complex system organized in a decentralized structure composed of three levels. At the highest level of complexity we have a traffic region, considered as a part of a citys road infrastructure. At the next level, the region is decomposed into traffic objects represented by road sections and intersections. At the last level of complexity we have the road segments that are parts of a road section. The traffic is considered in a macroscopic representation and the modeling for the different components of the road region is achieved by employing techniques from fluid mechanics. The results are verified in simulation, based on traffic data.

Compartmental networks approach on urban traffic control

Urban traffic networks play an important role in the modern society. Modeling and control of metropolitan traffic is a complex problem and a lot of research efforts are made towards solving this problem. We consider a macroscopic approach of the traffic and we propose a new solution for modeling and control based on mechanism offered by compartmental networks and positive systems theory. The traffic congestion is avoided by use of a nonlinear compartmental controller. A number of case studies and their results are presented.

Admissible modeling disturbances for the control of a road sector

Usually when modeling a real process the effects of nonlinearities must be taken into account. A road traffic sector is one such process. Considering it from a macroscopic point of view, the sector is modeled both as a nonlinear and as a linearized process. An 110 model is proposed and a RST control algorithm is used to ensure the imposed performances for the closed loop system. The robustness of the system is obtained by determining the modulus margin and based upon it the admissible nonlinearities are analyzed in order to determine the limits between which the system maintains its stability.

Advanced control and optimization for thermo-energetic installations

A new methodology for the design of control systems for real time applications is presented. It is proposed to have a first stage of assisted design of nominal systems with simulation based verification of the achieved performances. The control algorithm will then be improved by robustification to reproduce as good as possible the performances obtained in simulation on the real time process. Finally, a supervisor is implemented to optimize the thermo-energetic process and to compute the best choice for the reference signals.