Aleksandar Rakic

Early fault detection and isolation in coal mills based on self-organizing maps

Classical approaches to the fault detection and isolation usually require extensive plant-modeling and statistical analysis of the measured signals and their residuals versus the developed model. In this paper, alternative simple model-free approach is proposed. Real-time data are preprocessed and self-organizing map is trained and used for the reliable isolation of the most frequent mill fault — output fuel-mixture drop due to the coal-stuck in the input bunker. Proposed approach is successfully verified on the real-time data-sets from the coal mills in thermal power plant “Nikola Tesla B”, Serbia.

Simple fuzzy solution for quadrotor attitude control

This paper presents new computationally efficient and easy to tune attitude control of a four rotor helicopter called quadrotor. Short description of quadrotor concept and control framework is given and the segment of inner attitude control of the aircraft is addressed. Proposed fuzzy proportional - differential (PD) control structure utilizes Sugeno fuzzy engine, with simple intuitive set of rules and set of physically meaningful parameters. The verification of proposed solution was done via simulation and comparison with existing backstepping based attitude control system.

Interdisciplinary Project Bridges the Gap in Electrical Engineers' Knowledge of Modern Control Applications

The aim of this paper is to describe an interdisciplinary student project which combines knowledge and skills from power electronics, control algorithms and digital design. The suggested approach is a group-based project that enables students to develop interdisciplinary knowledge as well as interpersonal and communication skills needed to work in todays multidisciplinary field of modern control applications.

Induction motor torque control in field weakening regime by voltage angle control

This paper presents a novel algorithm for the induction motor torque control in field weakening region. Proposed method insures maximum DC bus utilization and offers DTC performance through the stator voltage angle control. The algorithm is simple, without the outer flux loop nor the inner current loop. Dynamic response is preserved over wide speed range by means of gain-scheduling. The paper comprises the implementation details and experimental results.

Stator Voltage Vector Direct Torque Control of induction machine

This paper presents improved high speed Induction Motor (IM) torque control algorithm that achieves full inverter voltage utilization in field weakening and smooth speed transition from the base speed to high speed region. The Stator Voltage Vector Direct Torque Control (SVV DTC) algorithm simultaneously controls IM torque and rotor flux by regulating stator voltage angle and amplitude in the base speed region, and by regulating only the voltage angle in the field weakening region. Fast torque dynamics is obtained in wide speed range by using torque regulator with variable gain. The main benefit from the proposed algorithm is in field weakening range, in which the machine is supplied by maximal available voltage without outer flux trajectory reference. This approach enables full utilisation of both magnetic material of the machine and power capabilities of the inverter, which is not possible with traditional vector drives with current regulators. Transition from base speed region to field weakening is obtained by simply limiting the output voltage. Experimental results gathered on the high speed low cost IM drive confirm the effectiveness of the proposed SVV DTC approach.

Architectural approach to cope with network-induced problems in network control systems design

Abstract In the field of process control engineering, network-based systems enable extensive, flexible and scalable applications in industrial automation and control. However, network-induced problems are influencing the stability and performance and they are introducing constraints in the system design and operation. While most of the existing design methodologies are searching for the specific solution within the domain of the control theory, we propose the comprehensive architectural approach that addresses wide range of the network-related issues and copes with them in the effective way. Presented solution combines several architectural styles encapsulating the actuating, sensing and control functionality into the unified service-oriented components, while the data transport is supported through event-triggered distributed middleware components. Given architectural approach decouples the design of process control functionality from the properties of the control network infrastructure. The effectiveness of the proposed solution is verified through the analysis of the system operation in the given case-study.

Cyber Security Against Denial of Service of Attacks on Load Frequency Control of Multi-area Power Systems

While open communication infrastructures are embedded into multi-area power systems to support data transmittion, it make communication channels vulnerable to cyber attacks, reliability of power systems is affected. This paper studies the load frequency control (LFC) of multi-area power systems under DoS attacks. The state space model of power systems under DoS attacks is formulated, where event-triggered control scheme is integrated for the multi-area power systems under DoS attacks. By utilizing average dwell time design approach, exponential stability and \(L_2\)-gain of the multi-area power systems can be obtained for event-triggered LFC scheme under DoS attacks, if choosing an unavailability rate of communication channels for DoS attacks properly. Finally, the example shows that the convergences of frequency deviation of three-area power systems are compared under different DoS attack scenarios, when the proportion of the total time of DoS attacks can obtain the result properly.

Novel AMI architecture for real-time Smart Metering

The proposed architecture of the Advanced Metering Infrastructure (AMI), presented in this paper, introduces new hierarchical layers in the Smart Metering System. This implies distributed implementation of the traditional data-concentrator functions as basic functionalities and advanced Smart Metering functions in the form of Transformer Station Data Concentrator (TSC) and Local Metering Concentrator (LMC) system components. This concept provides the support for real-time applications, services and algorithms. The introduction of LMC layer in the form of independent embedded system component enables on one side the real time two way command-response communication mechanism with the higher layers of AMI, while on the other side performs communication with locally connected electrical meters over standard serial communication interface. The proposed concept improves the communication dynamics with the traditional and smart electrical meters, while providing support for more complex AMI network topology. The introduced LMC layer enables real-time load profiling, energy diagnostic, and simple integration with energy management system. Interoperability, scalability and modularity are in the line with general smart grid concepts.

Variable sample rate acquisiton platform

This paper presents a solution for hardware platform for signal acquisition with variable sampling rate. The system is designed to work in noisy environment near power distribution grid with variable frequency.

Space Vector Representation of Induction Motor Model in Field Weakening Regime

In this paper a space vector model of an induction motor in a field weakening regime is detailed. Stator and rotor flux space vector trajectories are shown in a complex plane for the case when the machine is supplied by limited voltage.