Mario Vasak

University of Zagreb, Zagrabia, Grad Zagreb, Croatia

Are you Mario Vasak?

Claim your profile

Publications (45)14.75 Total impact

  • A. Starcic · A. Martincevic · M. Vasak
    [Show abstract] [Hide abstract] ABSTRACT: This paper investigates estimation of disturbance heat flux in a zone using autoregressive integrated moving average model with exogenous inputs (ARIMAX). The model is derived from simple continuous zone model. Data used for estimation is obtained from building simulation in IDA ICE. The procedure exhibits fast detection of the disturbance heat flux which opens several possible applications in energy-efficient buildings control, like possibility of instantaneous on-off reaction on the zone actuators or better building climate prediction for predictive control.
    No preview · Article · Jun 2015
  • Marko Gulin · Jadranko Matusko · Mario Vasak
    [Show abstract] [Hide abstract] ABSTRACT: In this paper we present power flow optimization of a residential DC microgrid that consists of photovoltaic array, batteries stack and fuel cells stack with electrolyser, and is connected to the grid via bidirectional power converter. The optimization problem aims to minimize microgrid operating costs and is formulated using a linear program that takes into account the storages charge and discharge efficiency. To account for power predictions uncertainty, optimization problem is defined in a stochastic framework by using chance constraints. Since we assume that the error in realization of power predictions will be compensated by utility grid, chance constraints are defined for power exchange between the microgrid and the utility grid. Finally, we investigate a stochastic model predictive control for the closed-loop power management in the microgrid. Performance verification of the proposed approach is performed on simulations for two-month period.
    No preview · Article · Jun 2015
  • Source
    Mladen Dalto · Jadranko Matusko · Mario Vasak
    [Show abstract] [Hide abstract] ABSTRACT: The aim of this paper is to present input variable selection algorithm and deep neural networks application to ultra-short-term wind prediction. Shallow and deep neural networks coupled with input variable selection algorithm are compared on the ultra-short-term wind prediction task for a set of different locations. Results show that carefully selected deep neural networks outperform shallow ones. Input variable selection use reduces the neural network complexity and simplifies deep neural network training.
    Full-text · Article · Jun 2015
  • M. Vasak · M. Baotic · N. Peric · M. Bago
    [Show abstract] [Hide abstract] ABSTRACT: The problem of energy-optimal control for rail vehicles under constraints and fixed arrival time is considered. The nonlinear train model with quadratic overall resistance force is approximated with a piecewise affine model. The optimal control law - train traction/braking force as a function of train speed and traversed path - is pre-computed off-line with the dynamic programming procedure. The optimal controller performance is verified in simulations.
    No preview · Article · Mar 2015
  • [Show abstract] [Hide abstract] ABSTRACT: The improvement of the building sector energy efficiency becomes crucially important to attain a balance in many sectors. Reduction of the energy consumption in buildings by using model predictive control strategies is recognized as one of the essential solutions to achieve considerable energy savings. Due to the nature of thermodynamic processes in buildings the underlying models are mostly nonlinear and of high order. In this work Constrained Unscented Kalman Filter is employed to obtain a linear low order model of a large public building applicable for the predictive control. Through the comparison of results with the data generated by highly accurate building simulation software IDA Indoor Climate and Energy (IDA-ICE), it has been shown that the first order linear model for each zone, with separated nonlinearities related to the solar radiation effects, is sufficient to capture the main dynamics of the observed building.
    No preview · Article · Jan 2015
  • [Show abstract] [Hide abstract] ABSTRACT: The paper concerns with improving the field-oriented controlled induction machine torque and speed accuracy and performance from a control theory viewpoint. A simple model is derived that includes first harmonics of real machine phenomena such as inherent asymmetries, saliencies and anisotropy. We design adaptive proportional-integral controllers based on that model to achieve best possible torque performance. The control algorithm is conceived as an extension of conventional induction machine rotor field-oriented control. The usual Kalman filter approach for state estimation required by field-oriented control is extended to identify the contribution and location of asymmetries in a dual filter approach with unscented Kalman filter. Simulation results are obtained for a 5.5 kW machine with modeled stator and rotor anisotropy. Undesired effects like torque and speed pulsations are greatly reduced with the proposed controller.
    No preview · Article · Dec 2014
  • [Show abstract] [Hide abstract] ABSTRACT: The paper concerns with increasing wind turbine availability by avoiding the shut-down under generator stator insulation degradation. Field-oriented control algorithm is derived for the wound rotor synchronous generators in wind turbine and further extended to achieve safe operation in the stator fault presence. The fault-tolerant control restricts the magnetic flux time derivative and thereby the stator voltage, which is the main cause for rapid fault development in the degraded stator insulation conditions. The method ensures maximum power production and wide wind turbine operation for different developed fault stages. The algorithm is set for the case of 700 kW wind turbine with direct-drive wound rotor synchronous generator but can be applied for any other type of generator as well.
    No preview · Article · Nov 2014
  • Marko Gulin · Mario Vasak · Tomislav Pavlovic
    [Show abstract] [Hide abstract] ABSTRACT: Residential microgrids are mainly based on renewable energy sources (i.e., mainly on photovoltaic panels), energy storage systems (that enable time-shift between production and consumption), and on power converters representing control points that by proper operation ensure overall system stability and quality of power supply. For optimal techno-economical microgrid operation, i.e. microgrid voltage level control and power flow management, the models of microgrid components involved must be known. In this paper we verify a single-diode five-parameters equivalent electrical model of a photovoltaic array by experiments, and based on the verified model propose a new power production model suitable for microgrid power flow optimization applications. Microgrid components are connected to a common power link via an appropriate power converter in order to ensure components' maximum efficiency and the overall system stability. Simulation of DC/DC power converters on a switching level requires significant computational efforts due to high switching frequencies (10-100 kHz). However, it is shown that dynamical behaviour of a DC/DC power converter in closed control loop can be replaced by a PT2 dynamic element with good approximation accuracy. All equivalent electrical models presented in this paper are implemented in the professional simulation platform for power electronic systems Plexim PLECS.
    No preview · Conference Paper · Sep 2014
  • Marko Gulin · Mario Vasak · Tomislav Pavlovic
    [Show abstract] [Hide abstract] ABSTRACT: In majority of cases residential microgrids are constituted of renewable energy sources, energy storage systems, and of power converters representing control points that by proper operation assure overall system stability and quality of power supply. In this paper we present simulation based analysis of dynamical behaviour of a residential DC microgrid laboratory setup in distributed and centralized voltage control configurations. It is shown that these control configurations have several inherent limitations, like overload of microgrid components during rapid load changes which can affect components lifetime. In order to overcome such limitations, the cause of such behaviour is assessed and control concepts to overcome that are proposed. Microgrid is simulated on electrical level using equivalent electrical models of all components involved: photovoltaic array, electrochemical batteries, fuel cells stack and power converters.
    No preview · Conference Paper · Jun 2014
  • [Show abstract] [Hide abstract] ABSTRACT: Microgrid is defined as a cluster of distributed generation sources, storages and loads that cooperate together in order to improve power supply reliability and overall power system stability. Short-term power production and load profile prediction is very important for power flow optimization in a microgrid, thus enhancing the management of distributed generation sources and storages in order to improve the microgrid reliability, as well as the economics of energy trade with electricity markets. However, short-term load prediction is a complex procedure, mainly because of the highly nonsmooth and nonlinear behaviour of the load time series. In this paper we develop and verify a neural-network-based short-term load profile prediction model. Neural network inputs are lagged load data, as well as meteorological and time data, while neural network output is load at the particular moment. Neural network training and validation is performed on load data recorded at University of Zagreb Faculty of Electrical Engineering and Computing, and on meteorological data obtained from Meteorological and Hydrological Service of Croatia, in period 2011-2013.
    No preview · Conference Paper · May 2014
  • [Show abstract] [Hide abstract] ABSTRACT: Renewable energy sources bring to fore the significance of power electronics and corresponding control. The topic of wind turbine fault-tolerant control has emerged as a promising line of research for improving their market competence. The paper presents summarized control of wind turbine converter with a fault-tolerant control extension for generator electromechanical faults, based on a proper modulation of the stator magnetic flux. The modulation also introduces hard periodic oscillations in the stator currents and generator output power. Simulation results obtained with a Matlab-Simplorer co-simulation and detailed power converter model show that DC-link capacitor and corresponding control algorithm act like an energy buffer that suppress the influence of generator stator oscillations on currents passed to the grid and maintain the high quality of power production. Following from the results, the generator and grid side controls are justified to be considered as two independent problems.
    No preview · Conference Paper · Apr 2014
  • [Show abstract] [Hide abstract] ABSTRACT: Electrical machines in the high-voltage class are usually designed with open stator slots. This wide open slots cause an increase of higher order harmonics, vibrations, noise, and temperature; thus, the machine efficiency is decreased. To counteract this disadvantage, magnetic slot wedges are applied. Due to the impact of high magnetic and mechanical forces, these wedges can fall out and may cause further serious damages. Up to now, reliable detection methods for single missing slot wedges are coupled with a disassembling of parts of the machine. In this paper, a method is investigated which provides the possibility of detection, based on the measurement of electrical terminal quantities only.
    No preview · Article · Mar 2014 · IEEE Transactions on Industry Applications
  • [Show abstract] [Hide abstract] ABSTRACT: Before applying current-signature-analysis-based monitoring methods, it is necessary to thoroughly analyze the existence of the various harmonics on healthy machines. As such an analysis is only done in very few papers, the objective of this paper is to make a clear and rigorous characterization and classification of the harmonics present in a healthy cage rotor induction motor spectrum as a starting point for diagnosis. Magnetomotive force space harmonics, slot permeance harmonics, and saturation of main magnetic flux path through the virtual air-gap permeance variation are taken into analytical consideration. General rules are introduced giving a connection between the number of stator slots, rotor bars, and pole pairs and the existence of rotor slot harmonics as well as saturation-related harmonics in the current spectrum. For certain combinations of stator and rotor slots, saturation-related harmonics are shown to be most prominent in motors with a pole pair number of two or more. A comparison of predicted and measured current harmonics is given for several motors with different numbers of pole pairs, stator slots, and rotor bars.
    No preview · Article · Sep 2013 · IEEE Transactions on Industrial Electronics
  • [Show abstract] [Hide abstract] ABSTRACT: This paper concerns with wind turbine generator fault-tolerant control that avoids system shut-down and enables safe operation with less than nominal power production. We focus on generator stator isolation inter-turn fault that can be characterized before triggering the safety device and applied for any type of generator used in wind turbines. A low complexity model predictive tracking controller is proposed to achieve very accurate flux modulation that prevents the fault propagation while power delivery under fault is deteriorated as less as possible compared to healthy machine conditions. Presented fault-tolerant control strategy is developed taking into account its modular implementation and installation in available control systems of existing wind turbines to extend their life cycle and energy production. Simulation results for the case of a 700 kW wind turbine are presented.
    No preview · Conference Paper · Jan 2013
  • Source
    M. Vasak · A. Martincevic
    [Show abstract] [Hide abstract] ABSTRACT: Approximately 42% of total energy usage today is spent on the heating of the human environment. Energy consumption can be significantly reduced by proper design of the heating control system - performed studies suggest that in average about 12% reduction of energy usage can be achieved based on control system improvement. This paper describes an efficient control of a family house heating system in its very common configuration. We use model predictive control to fully take into account the prediction of disturbances (room temperatures), reference values of power per rooms and physical limitations of actuators. This will result in an optimal flow rate and temperature of medium required to meet all the demands. In this paper we will first present the model of a heating system, and from it derive model predictive control and verify it through simulation.
    Full-text · Conference Paper · Jan 2013
  • Source
    M. Vasak · A. Starcic
    [Show abstract] [Hide abstract] ABSTRACT: The largest energy consuming subsystem in a building is heating, ventilation and air conditioning (HVAC) system. It is thus essential to optimize the behavior of that subsystem in order to achieve energy savings. In this work focus is on efficient energy usage in a sample zone system consisting of two coupled zones. Ventilation, heating and cooling are done by air handling unit (AHU) and variable air volume (VAV) boxes. Optimization of the system means finding optimal control sequences for AHU and VAV boxes. The model of the coupled zones is linear, while the models of the AHU and VAV boxes are nonlinear. In order to find a suboptimal solution, which will keep zone temperature in given boundaries, while handling ventilation requirements, linear programming and linear zone system model are used to find appropriate cooling/heating powers, and then those powers are used to obtain control sequence for AHU and VAV boxes.
    Full-text · Conference Paper · Jan 2013
  • [Show abstract] [Hide abstract] ABSTRACT: The winding system of high voltage machines is usually composed of pre-formed coils. To facilitate the winding fitting process stator slots are usually wide opened. These wide opened slots are known to cause disturbances of the magnetic field distribution. Thus losses are increased and machine's efficiency is reduced. A common way to counteract this drawback is given by placing magnetic slot wedges in the slots. During operation the wedges are exposed to high magnetic and mechanical forces. As a consequence wedges can get loose and finally fall out into the air-gap. State-of-the-art missing slot wedge detection techniques deal with the drawback that the machine must be disassembled, what is usually very time consuming. In this paper a method is investigated which provides the possibility of detecting missing magnetic slot wedges based only on measurement of electrical quantities and without machine disassembling. The method is based on exploitation of machine reaction on transient voltage excitation. The resulting current response contains information on machine's magnetic state. This information is composed of several machine asymmetries including the fault (missing wedge) induced asymmetry. A specific signal processing chain provides a distinct separation of all asymmetry components and delivers a high sensitive fault indicator. Measurements for several fault cases are presented and discussed. A sensitivity analysis shows the high accuracy of the method and the ability to detect even partially missing slot wedges.
    No preview · Conference Paper · Oct 2012
  • [Show abstract] [Hide abstract] ABSTRACT: A method for dynamic modelling of induction machine with a doubly slotted air gap is proposed and implemented for the case of a cage induction motor. The described method is easily extensible to wound rotor machines. A numerical description of the air gap permeance is provided that takes into account a slotted stator and rotor structure as well as their mutual, time and space dependant positions as a function of rotor rotation. The multiple coupled circuit model approach is used with the modified winding function in order to calculate the inductance of all motor windings. The developed model is general in nature and could be used for the analysis of different dynamic regimes of induction machine, particularly different combinations of stator and rotor slot numbers. Model validation is provided by stator current spectrum analysis of a standard four pole induction motor with S=36 and R=32 slots. The experimental results presented clearly support these findings.
    No preview · Conference Paper · Sep 2012
  • [Show abstract] [Hide abstract] ABSTRACT: Electrical machines in the high voltage class are usually designed with open stator slots. This wide open slots cause an increase of higher order harmonics, vibrations, noise and temperature, thus the machines efficiency is decreased. To counteract this disadvantage magnetic slot wedges are applied. Due to the impact of high magnetic and mechanical forces these wedges can fall out and may cause further serious damages. Up to now reliable detection methods for single missing slot wedges are coupled with a disassembling of parts of the machine. In this paper an method is investigated which provides the possibility of detection, based on the measurement of electrical terminal quantities only.
    No preview · Conference Paper · Sep 2012
  • [Show abstract] [Hide abstract] ABSTRACT: Inverter fed induction machine drives have become one of the most important part in industrial applications. The devices are usually operated near or at the rated vales. Additionally high dynamic operations, overload cycles and the fast switching inverter are putting additional stress to all machines components. The stator winding system is one of the most critical components. Faults in the winding system of a machine can cause a dramatic damage due to their fast developing behavior. An early detection of winding faults can reduce these consequences. In this paper a method is presented to achieve reliable fault detection on stator winding faults focused on application at higher speed levels where immediate detection is imperative. Higher inverter modulation indexes provide the possibility to estimate the transient leakage inductance in each PWM modulation cycle. This parameter contains the information on machine asymmetries. Specific signal processing steps provide a high sensitive fault indicator.
    No preview · Conference Paper · Sep 2012