Oscillations in Inverter Fed Induction Motor Drives

Abstract :

Severe oscillations in the range of 1 to 100 Hz have been encountered in inverter fed induction motor drive systems, especially where there are no external damping loads, such as fan drives. These oscillations may damage the drive system or generate noise. It is found that the induction machine has two resonance frequencies. The damping of the first resonance is decreased with increased stator resistance, while the damping of the second resonance is increased with increased stator resistance. Simple mechanical models are presented which give physical insight into the reason for the oscillations, as well as suggestions of how to suppress them.

On the Verifiability of the Activated Sludge System Dynamics

Abstract :

Wastewater treatment processes are inherently dynamic because of variations in the influent flow rate, concentration, and composition. The adaptive behaviour of the microorganisms further emphasizes this fact. Mathematical models and computer simulations are essential to describe, predict, and control the complicated interactions of the processes. Any attempt to model all details of the various reaction mechanisms are, however, destined to fail due to lacking knowledge and the extreme complexity required for such models. A reduced order dynamic model for an activated sludge process performing carbonaceous removal, nitrification, and denitrification is presented herein. The identifiability of the model is investigated using both off-line and on-line methods and its dynamic behaviour is verified against simulations of a recognized model - the IAWPRC Activated Sludge Model No. 1. The required data for the identification algorithms is based on directly measurable real time data. The simplified model may serve as a tool for predicting the dynamic behaviour of an activated sludge process since the parameters under varying operating conditions can be tracked on-line. The model is aimed for operation and control purposes as an integral part of a hierarchical control structure.

Computer-supported Complexity Reduction in Process Control - A Cognitive Approach to User Interface Design

Abstract :

The design of the user interface is one of the most important aspects in the development of computer systems. Many results are available to guide this design in relation to the appearance of the interface, but not on how to support human cognitive capabilities in relation to the tasks to accomplish. This work presents a new approach to include cognitive factors in the design of user interfaces. Every human-computer interaction problem is considered as a cognition problem of dealing with an unknown, complex system. The user interface must act as complexity interface between the technical process and human cognitive capabilities. The exposition draws on current results reported in the literature and organizes them into a common, application-oriented frame. Many examples about how interfaces work in practice and unexpected results in real applications are used to support, qualify, or question the experimental and theoretical findings. The proposed methodological approach is used in a case-study evaluation of the design of the user interface for the monitoring and control system in a satellite ground control station.

Induction Machine Speed Estimation - Observations on Observers

Abstract :

This work focuses on observers estimating flux linkage and speed for induction machines, mainly in the low speed region. With speed estimation, sensorless control is possible, meaning that the speed of induction machines without mechanical speed sensors can be controlled. The observer based sensorless drive system has superior dynamic performance compared to a system with an open loop frequency inverter, yet it is neither more complex nor expensive. Using mechanical equivalent models of the induction machine and observers, an accurate flux observer working in the entire speed region of the induction machine is presented. The flux observer is expanded into a combined flux and speed observer, measuring only stator current and voltage. A method for sensorless control is proposed, analyzed and experimentally verified. Observer and controller calculations are performed by a digital signal processor.

Modelling Aspects of Wastewater Treatment Proccesses

Abstract

Wastewater treatment processes are inherently dynamic because of the large variations in the influent wastewater flow rate, concentration and composition. Moreover, these variations are to a large extent not possible to control. The adaptive behaviour of the involved microorganisms imposes further difficulties in terms of time-varying process parameters. Mathematical models and computer simulations are essential to describe, predict and control the complicated interactions of the processes. The number of reactions and organism species that are involved in the system may be very large. An accurate description of such systems can therefore result in highly complex models, which may not be very useful from a practical, operational point of view.

A reduced order dynamic model, describing an activated sludge process performing carbonaceous removal, nitrification and denitrification with reasonable accuracy, is presented. The main objective is to combine knowledge of the process dynamics with mathematical methods for estimation and identification. The identifiability of the model is investigated using both off-line and on-line methods, and its dynamic behaviour is validated by simulations of a recognized model. The information required by the identification algorithms is based on directly measurable real-time data. The simplified model may serve as a tool for predicting the dynamic behaviour of an activated sludge process, since the parameters can be tracked on-line during varying operating conditions. The model is aimed for operation and control purposes as an integral part of a hierarchical control structure.

The main objective of the work on settler modelling is to enlighten recent theoretical results. A new one-dimensional settler model is compared to a traditional layer model by means of numerical simulations. Emphasis is put on the numerical solution?s ability to approximate the analytical solution of the conservation law written as a non-linear partial differential equation. The new settler model is consistent in this respect. Several problems that occur when integrating a model of the biological reactor with a model of the settler are also discussed. In particular, the concentrations of the biological components of the particulate material are of importance for an accurate description of the sludge that is recycled to the biological reactor. Two one-dimensional algorithms have been evaluated. The first algorithm is commonly used and some of its inherent problems are discussed. The second algorithm is a new analytically derived method.

Few attempts have been made to take into account the influence of higher order organisms in biofilm systems when developing or applying mathematical models. This work describes a simplified modelling approach to include some possible effects of higher order organisms on nitrification, based on a proposed hypothesis of their oxygen consumption in the biofilm. Three different models are developed and investigated. Model simulations are validated using data from a laboratory experiment using continuous-flow suspended-carrier biofilm reactors, where the predators were selectively inhibited. The proposed models are capable of reproducing several of the observed effects. They are primarily aimed at capturing the steady-state behaviour of the biofilm but may also prove to be a useful basis for describing the dynamics.

Diagnosis Problems in Wastewater Settling

Abstract :

Solids-liquid separation is of crucial importance in wastewater treatment systems, in particular in activated sludge systems. This is true in a wide spectrum of operating conditions. During severe hydraulic loading the clarifier becomes the bottle-neck. Likewise, the long term development of the microbial species determines the floc structure, and this in turn is directly reflected in the separability of the sludge. Due to these reasons, it is essential to find ways to automatically detect any changes in the settling properties. This work demonstrates ways to use simple on-line measurements to give the operator a more elaborate picture of the process behaviour. Occasionally the suspended solids concentration may display both large amplitude and unnaturally fast changes. If the sensor calibration has been found satisfactory, the abnormal appearance has to be explained by the process itself. There may be three principal reasons for such a behaviour: severe hydraulic conditions, poor floc separation properties or a high sludge blanket. During these circumstances it is very informative to qualitatively examine the relationships between the suspended solids concentration variations and important influencing variables, such as the flow rates and the sludge blanket level. It may be difficult to find accurate dynamical models for such extreme behaviour. However, a gross measure of the process state is often sufficient information to the operator to initiate further off-line or laboratory analysis in order to find the underlying cause. Having found consistent data of flow rates and turbidity their dynamical relationships can be further analysed. Generally, there is an obvious dynamical causality between the hydraulic conditions, the sludge blanket level and the effluent suspended solids concentration. Two different approaches have been investigated: parameter estimation and rule based reasoning. In this kind of application parameter estimation in time series models is not generally successful. Their drawbacks are displayed and discussed. Instead of a quantitative approach a qualitative rule based ditto is suggested. Several advantages of rule based methods are demonstrated, for example their capability to deal with unexpected situations that always appear in on-line applications. The ability to cope with on-line data is one crucial factor when looking for a methodology for diagnosis in wastewater facilities. As in any parameter identification method, sufficient excitation of the system, e.g. by flow changes, is required. In other words, not all consistent data series are suitable for diagnosis. However, rule based methods may be applied to a significantly wider range of situations than parametric dittos. This is one important punch line of this work. Odd situations with almost momentary disturbances in the return sludge flow and effluent suspended solids concentration are displayed and explained. Their appearance initiate a discussion of ways for active diagnosis, where peak disturbances may be purposefully injected to the return sludge flow in order to validate the settler operational marginal.

Torque Control of a Novel Switched Reluctance Machine

Abstract :

This work is focused on torque control and performance analysis of two switched reluctance machines having six stator poles and four rotor poles. The first machine is a conventional design, built with a laminated iron core, whereas the second is a novel construction where the magnetic flux conductor consists of a soft magnetic composite material. A table based torque control scheme is developed which allows the user to specify the permissible torque ripple. The current reference value for all phases are written into a large memory, which is addressed by the torque reference, rotor angular position and rotor angular speed, all in digital form. The phase currents are controlled with hysteresis controllers, one for each phase. The updating frequency of the controller is in excess of 100 kHz, which makes it attractive for high speed applications. Models for the iron an copper losses as well as for the torque production capability of thw switched reluctance machine are developed. Both the theoretical analysis and experimental verification are performed on two switched relluctance machines rated 9.5 Nm at 4600 rpm. It is shown that the iron losses in both machines are almost the same, and that 70-75 % of the iron losses are confined to the stator. The iron losses greatly exceed the copper losses at the nominal operating point. Finally, it is concluded that increasing magnetic frequence is favourable if the iron core is made with the soft magnetic composite material tested.

Borrowing the Bat's Ear for Automation - Ultrasonic Measurements in an Industrial Environment

Abstract :

This work focuses on measurement principles based on ultrasound for applications in industrial automation. The measurements made are of geometric nature. The methods used are when possible based on experience and/or inspiration from the study of bats. A platform for ultrasonic measurements, that allows a VME-based host computer to transmit and receive signals, with a considerable flexibility has been built. Sensor units utilizing both piezoelectric and electrostatic transducers are used. The dimensions of the units are about 10x10x7 centimeters. Frequencies in the range 40-200 kHz are used. The presented measurements include distance, flat surface spatial angle, object localization, object orientation, and object recognition. To make the methods robust matched filters are used in several of the measurements. Frequency sweeps are used to make the ultrasonic measurements more robust to various environmental parameters. Methods that utilize the new information provided by the frequency sweeps are also suggested.

Some Control Applications in Electric Power Systems

Abstract :

This thesis presents some control applications in electric power systems. The work consists of the four parts: interaction between DC cables and ship steering autopilots; frequency estimation; computer relaying; field tests using load switching to damp power oscillations.

The first three parts are based on four papers that are given in Appendix A-D. The thesis presents supplementary text to these papers. The last chapter is more self contained and presents field tests from a small hydro power station. These tests show that load switching is an excellent method to damp power oscillations. A proof used in stochastic analysis of fault locators is presented in Appendix E.

The first paper presents novel results on interaction of DC cables and ship steering autopilots. Simulations with a non-linear model show that a ship using steering autopilot with a magnetic compass can be captured by a DC cable, i.e, the ship tracks the cable. A full scale experimental study has been performed at the Kontiskan HVDC link. Good agreement was obtained between measurements and computer sinulations.

The second paper presents a method for frequency estimation in power systems by demodulation of two complex signals.

The third and fourth paper presents some improvements of the differential equation algorithm (DEA) that can be used for transmission line protection. The third paper uses the determinant to explain why isolated estimates from the DEA can be very poor. A median filter is proposed to reject these poor estimates. The fourth paper suggests the followinmg improvements to the DEA: fault classification; intermediate filtering and a new algorithm for three phase faults. Simulation results indicate a nominal operation time of 5-7 ms for three phase faults and slightly longer for other fault types.

Power System Damping - Structural Aspects of Controlling Active Power

Abstract :

Environmental and economical aspects make it difficult to build new power lines and to reinforce existing ones. The continued growth in demand for electric power must therefore to a great extent be met by increased loading of available lines. A consequence is that power system damping is reduced, leading to a risk of poorly damped power oscillations between the generators. This thesis proposes the use of controlled active loads to increase damping of such electro-mechanical oscillations. The focus is on structural aspects of controller interaction and of sensor and actuator placement.

On-off control based on machine frequency in a single machine infinite bus system is analysed using energy function analysis and phase plane plots. An on-off controller with estimated machine frequency as input has been implemented. At a field test it damped oscillations of a 0.9 MW hydro power generator by controlling a 20 kW load.

The linear analysis uses two power system models with three and twenty-three machines respectively. Each damper has active power as output and local bus frequency or machine frequency as input. The power system simulator EUROSTAG is used both for generation of the linearized models and for time simulations.

Measures of active power mode controllability and phase angle mode observability are obtained from the eigenvectors of the differential-algebraic models. The geographical variation in the network of these quantities is illustrated using the resemblance to bending modes of flexible mechanical structures. Eigenvalue sensitivities are used to determine suitable damper locations.

A spring-mass equivalent to an inter-area mode provides analytical expressions, that together with the concept of impedance matching explain the structural behaviour of the power systems. For large gains this is investigated using root locus plots. The effect of using two dampers is studied. For the three machine system this is done for all combinations of the two gains in a certain range. In the twenty-three machine case one gain takes only two values as the other is varied.

Direct Load Control - Power Peak Shaving Applied to a Foundry

Coordinated Tap Changer Control - Theory and Practice

Abstract :

This thesis deals with oscillation phenomena related to the discrete nature of tap changers. Two fundamentally different phenomena are investigated:

• oscillations due to interaction among the local control systems of cascaded tap changers
• self-sustained oscillations due to interaction between the tap changer control system dead-band and load dynamics

• new tuning for the existing (local) controllers
• a centralised optimal controller
• a centralised fuzzy rule-based controller

The second phenomenon is of importance for the analysis of oscillatory voltage collapse. Using a small example system, it is shown that the results from analysis based on a continuous state OLTC model are unreliable, since the same system with a detailed OLTC model exhibits a limit cycle related to the control system dead-band that will arrest the oscillatory voltage instability predicted by small disturbance analysis. The key parameters for the occurrence of limit cycles are identified as the system load level, degree of reactive compensation and the load voltage dependency. Adjusting OLTC control system parameters such as time delays or dead-band size are shown to have no effect on the existence of these limit cycles in certain loading conditions.

The Back-to-back converter - theory and design

Abstract :

No abstract available. See full document.

Monitoring Wastewater Treatment Systems

Abstract :

This work considers various techniques to extract information from the vast amount of data collected at a modern wastewater treatment plant. If the information extracted is to be considered reliable is highly dependent on the data screening. Data screening includes validation and quality improvement of data. Adequate methods for validation, noise reduction and other forms of quality improvements of wastewater treatment data are discussed. In order to detect deviations and disturbances, the measurement variables can be investigated individually or many variables simultaneously. Single variable detection involves investigation of the basic signal characteristics such as amplitude, mean and spread. Usable methods are discussed and examples are given. In order to detect synergetic effects, techniques capable of investigating several variables simultaneously, are needed. Multivariate statistics based methods, such as principal component analysis (PCA), principal component regression (PCR) and projection to latent structures (PLS), are considered and their applicability discussed. Some possibilities to adapt the methods to the dynamic situation in a wastewater treatment plant are also outlined.

A Power flow Control Strategy to Minimize Energy Losses in Hybrid Electric Vehicles

Abstract :

This work considers control of power ow and energy storage in a Hybrid Electric Vehicle (HEV). The performance of a HEV, such as fuel consumption and emissions, depends not only on the individual components but also on how the components are controlled. Control in this case is to determine operating points for the components during driving, while control on the machine level, such as speed and torque control, is not an issue. A model of a series HEV has been developed. A number of di erent performance indices have been applied to the model and evaluated using methods from optimal control. These methods allow us to dene simple speci cations such as minimizing fuel consumption and to put constraints on emissions. The resulting controller is impossible to implement since it is open loop and needs information about the future. A controller that captures the behaviour of the optimal controller has been synthesised and tried on the model with satisfying result.

The developed controller has also been implemented on a full scale exper- imental platform, similar to the hybrid busses in Stockholm. The experi- mental and theoretical results show good agreement. A reduction in fuel- consumption of about 10% seems possible when the results are compared with the busses in Stockholm.

Advanced Control of Active Filters in a Battery Charger Application

Abstract :

In this thesis, a high performance battery charger for electric vehicles (EVs) is investigated. By including active power line conditioning capabilities in the battery charger, a viable concept for a fast charging infrastructure is obtained, beneficial both to the EV users and the power distributors. The thesis contains discussions on modelling aspects and design considerations for the proposed battery charger, based on a carrier wave modulated self commutated 2-level voltage source converter topology. Furthermore, model based controller synthesis is employed, and thorough analysis of the controller characteristics is given. Emphasis is put on the active filtering performance of the battery charger. The weaknesses of the model based control system in active filter applications are revealed, where especially the inherent phase deviation of the control system and the sensitivity to system parameters deteriorates the performance of the active filter. In order to overcome the deteriorating properties of the model based controller, a controller structure for active filters based on several integrators in multiple reference frames is proposed. It is shown, both theoretically and experimentally, that the proposed controller exhibits low sensitivity to system parameters and provides for complete compensation of the inherent phase deviation. The result is excellent conditioning performance of the proposed active filter controller in steady state.

Quasi Resonant DC Link Converters - Analysis and Design for a battery Charger Application

Abstract :

Environmental aspects have during the last years made electric vehicles an interesting competitor to the present internal combustion engine driven vehicles. For a broad introduction of pure electric vehicles, a battery charging infrastructure is deemed necessary. However, to build and maintain such an infrastructure is costly. Active power line conditioning capabilities could be included in the battery charger, making the infrastructure advantageous from the distribution network point of view. Another option is that the battery charger could be able to support the grid with peak power during periods of stability problems or emergency situations. This means that energy is borrowed from the batteries of vehicles connected to a charger. The price for energy supplied by the batteries is likely to be several times higher than the normal electric energy price, due to the wear costs of the batteries. Therefore, the battery charger losses also represents a high cost. It is often stated that resonant converters have a high efficiency compared to hard switched. Since carrier based pulse width modulation is employed, quasi resonant DC link converters are of interest. Four of the most promising quasi resonant DC link topologies reported in the literature are compared. A fair comparison is obtained by designing them to meet certain common design criteria, in this case the duration of the zero voltage interval and the maximum output voltage time derivative. The derivation of the design expressions are given, and also the simulation results, by means of efficiency. A 10 kW battery charger, equipped with one of the quasi resonant DC links investigated is implemented. A hard switched battery charger with the same rating is also tested to compare the measured efficiency with the simulated. Both the simulations and measurements shows that the efficiency decrease for quasi resonant battery chargers compared to the hard switched case. Furthermore, low frequency harmonics appear in the battery charger input and output currents. However, full control of the output voltage derivatives is obtained.

Optimization of a Servo Motor for an Industrial Robot Application

Abstract :

Automated Manufacture of Fertilizing Granules from Burnt Wood Ash

Abstract :

This work considers control of an ash transformation process, which transforms wood ash produced at district heating plants into fertilizing granules. The manufactured granules are recycled back to the forest grounds, as a fertilizer, or as a tool to reduce the acidification in the forest soil at the spreading area. Other areas of application are, for example, structural fill and substitute for cement in ready-mix concrete.

The ash transformation process includes mixing, size reduction, granulation, drying, sorting and packing. Furthermore, the transformation process uses ETEC-dolomite that acts as a binding agent and therefore improves the strength of the produced granules. Due to the dolomite, the lime effect of the produced fertilizers is increased, whereas the leaching speed of the wood ash included heavy-metals is decreased. A robust machine is developed and controlled by an industrial control system in order to enable continuous and automatic manufacture. At present, the units for mixing and size reduction are fully implemented, built to comply with the industrial requirements for continuous operation. The remaining stages are controlled to a certain extent, but are still based on the earlier prototype.?

Mixing ash/dolomite/water in order to obtain granular material is one method to stabilize wood ashes. The main problem is predicting the quantity of water to be added since the necessary amount varies with the wood ash quality. The implemented controller is therefore able to determine this optimal amount without any measure of the wood ash quality, as for example the ash carbon content. However, the produced granules do not benefit from high carbon content. Therefore, two potential on-line methods for carbons in ash monitoring are presented but not implemented due to financial reasons.?

Coordinated Voltage Control in Electric Power Systems

Abstract :

This thesis deals with various aspects of coordinated voltage control in electric power systems. It consists of an introduction and six papers. The introduction contains a summary of the thesis and tutorials on voltage control and the search methods that have been applied.

The first part deals with the coordination of cascaded tap changers in radial distribution feeders. Poor coordination causes unnecessary operations of the tap changers and consequently unnecessary wear as well as poor voltage quality. A tuning rule for the conventional (local) controllers as well as two new (cen-tralized) schemes are proposed: firstly, an optimal controller based on on-line optimization and secondly, a rule based controller based on a fuzzy inference system. Results from simulations based on load patterns recorded during dif-ferent seasons indicate that when tuned according to our recommendation, tap changers perform some 10 % fewer operations compared to when the present tuning is in use. The additional reductions with the fuzzy-rule based and op-timal controllers are some 36 % and 45 %, respectively. A prototype of the fuzzy-rule based controller has been installed and successfully tested in a distri-bution feeder in the south of Sweden. The prototype was also used to validate the simulation results presented in the papers.

Another control interaction problem studied is a phenomenon of limit cy-cles due to load-tap changer interaction. The role of the OLTC deadband in creating a limit cycle phenomenon is addressed. Analytical insights into the be-haviour of power systems with continuous and discrete tap models are given using eigenvalue analysis, the describing function method and time simula-tion. The conditions for occurrence of limit cycles and the identification of key parameters influencing their characteristics are other important aspects of this investigation.

The second part of the thesis describes a coordinated system protection scheme (SPS) against voltage collapse based on model predictive control. The combinatorial optimization problem that results from the control problem for-mulation is solved by a search method adapted from the research on artificial intelligence. Using predictions of the system?s future behaviour, the SPS opti-mally coordinates dissimilar and discrete controls such as generator, tap changer and load shedding controls in presence of load dynamics and constraints on controls as well as voltages and currents in the network. The response with the coordinated SPS is compared to a local scheme by simulation of some critical contingencies in the CIGRE Nordic 32 test system. The simulation results in-dicate that local undervoltage load shedding is near-optimal when coordination with other emergency controls is not necessary. When also other controls such as generator voltage setpoints are considered, the coordinated scheme reduces the amount of load shedding necessary to restore voltage stability by some 35 %. The reduction in the amount of load shed is mainly due to improved utilization of remote generators with excess reactive capacity to spare.

Modelling of power systems has also been addressed. The third part de-scribes a freely available power system library called ObjectStab intended for voltage and transient stability analysis and simulation written in Modelica?a general-purpose object-oriented modelling language. All component models are transparent and can be modified or extended. Power system topology and parameter data are entered in one-line diagram form using a graphical editor. The component library has been validated using comparative simulations with Eurostag.

Observations on Electric Hybrid Bus Design

Abstract :

In this work a simulation model is for two completely different commercial series hybrid buses is present. The simulation model concerns all the major power flow, temperatures and Internal Combustion Engine (ICE) parameters in the busses. The model is verified through measurements on hybrid buses in Malmö and Stockholm and test bench driving of the ICE. Proposal improvements of the charging strategy in the busses and regarding size of the different components in the drive train. The charging strategy means the way the ICE is driven. The aim is minimize the fuel consumption and emissions and improvement of the energy handling system. Practical experience of handling the vehicles is also part of the work.

A Chemomtric Approach to Process Monitoring and Control - With Applications to Wastewater Treatment Operation

Abstract :

In this work, various aspects of multivariate monitoring and control of wastewater treatment operation are discussed. A number of important difficulties face operators and process engineers when handling online measurements from wastewater treatment processes. These include, for instance, a high number of correlated measurement variables, non-stationarities, nonlinearities and multiscale process behaviour. A systematic way to handle and analyse data is needed to effectively extract relevant information for monitoring and control. In this work, a chemometric approach is taken. Principal component analysis (PCA) is used to reduce both the dimensionality of the problem and the noise level in data. However, it is shown that basic PCA is not sufficient to describe the process adequately. There are mainly two reasons for this. First, the processes display a non-stationary behaviour due to the diurnal, weekly and seasonal variations in the composition of the wastewater. Second, disturbances and events occur at different time scales making basic PCA less suitable.

The problem of non-stationary data is overcome using adaptive PCA in terms of updating of the scale parameters as well as the covariance structure. It is shown that adaptive PCA significantly improves the monitoring results as the model adapts to new process conditions without losing its ability to detect deviating process behaviour. To solve the problem of disturbances that occur at different time scales multiscale PCA is used. Multiscale PCA is a combination of multiresolution analysis and PCA. Measurement signals are decomposed into several time scales, and PCA models at each scale are identified. By doing so, the sensitivity to small process deviations that otherwise are obstructed by the diurnal variation is considerably increased. By omitting the lowest time scale from the analysis, the remaining time scales will inherently be (practically) stationary since this corresponds to using a highpass filtered version of the data. Another solution, where the PCA models at each scale are made adaptive is also presented.

Using the monitoring results to adjust the process in a supervisory control manner is discussed. Two different methods are presented. The first is based on a multistep procedure. The current operational state is detected and classified using clustering in the principal component space. This information is used to determine appropriate setpoints for local controllers so that the process returns to what is considered normal operation. In the setpoint determination step, both static and dynamic models are used. The dynamic models are used within the framework of model predictive control (MPC). The multistep approach is best suited for extreme event control, since nonlinear and discrete control actions easily can be incorporated. The second method to integrate monitoring and control is based on PCA. Here, the inverse PCA model is used to directly calculate appropriate setpoints for the local controllers so that the process can be controlled to attain specified output requirements. The controller can be seen as a multivariate feedback controller implemented on top of the local control system. It is shown by simulation studies that both methods for supervisory control can successfully be used to control the process according to the control objectives.

Acoustic Noise in Vector Controlled Induction Motor Drives

Abstract :

This thesis deals with the analysis and prediction of the acoustic noise emitted from vector controlled induction motor drives. The effect of some structural changes in the stator, introduced with the aim of reducing the noise emissions, is also discussed.

Noise pollution in the environment and in the workplace can cause discom- fort, a change in performance and even physical damage. Electric machines, and more in particular induction motors, are nding application in new envi- ronments where lower noise levels are demanded. It is, therefore, imperative that the designs should satisfy certain criteria in terms of noise generation and transmission. Vector control is one of the most commonly implemented strate- gies in variable speed drives, and hence the importance of understanding the characteristics of the noise generation when this technique is used.

A digital drive system has been developed for a 2.2 kW induction motor, and its dynamic capabilities demonstrated for a wide range of the frequency spectrum. This tool has been used for the experimental evaluation of the noise emissions when the ux and/or the torque are modulated with high frequency noise signals. The results showed that the noise emissions were higher when the ux was modulated compared to the torque, although the differences were considerably reduced when the machine was loaded. It was also observed that the noise emissions were decreased importantly at load. Sound pressure and sound intensity measurements have been conducted with the rotor stationary and rotating at low speed, showing that the most proper way to quantify the noise emissions from electric machinery is to measure the sound power through sound intensity measurements.

A method for the prediction of the noise emissions has been proposed, based in the interactive use of mechanical and electromagnetic nite element analyses, together with an acoustic boundary element method. The results show that the accuracy of the noise prediction depends on the proper calculation of the modes of vibration in the structural analysis, as well as a suitable selection of the material damping. The effect of skewing also needs to be modelled in order to account for high order rotor harmonics. The inuence of a 10% increase in the saturated part of the iron BH curve is translated into an increase of 3 dB or less in the sound power level, depending on the frequency.

The effect of the introduction of air gap layers around the outer part of the stator core in order to interfere with the natural transmission path of the vibrations has also been analyzed. When the gaps are lled with air the acoustic behaviour is not improved since the reduction of the stiffness in the outer core actually increased its sensitivity to the vibrations.

Realising Full-Scale Control in Wastewater Treatment Systems Using In Situ Nutrient Sensors

Abstract :

A major change in paradigm is taking place in the operation of wastewater treatment plants as automatic process control is becoming feasible. This change is due to a number of different reasons, not least the development of online nutrient sensors, which measure the key parameters in the biological nutrient removal processes, i.e. ammonium, nitrate and phosphate.

This thesis is about realising full-scale control in wastewater treatment systems using in situ nutrient sensors. The main conclusion of the work is that it is possible to significantly improve the operational performance in full-scale plants by means of relatively simple control structures and controllers based on in situ nutrient sensors. The in situ location should be emphasised as this results in short dead time, hence making simple feedback loops based on proportional and integral actions effective means to control the processes.

This conclusion has been reached based on full-scale experiments, where various controllers and control structures for the biological removal of nitrogen and the chemical removal of phosphorous have been tested. The full-scale experiments have shown that it is possible to provide significant savings in energy consumption and precipitation chemicals consumption, reduction in sludge production and improvement of the effluent water quality.

The conclusions are supported by model simulations using the COST benchmark simulation platform. The simulations are used for investigating issues regarding the interactions between the main control handles working in the medium time frame (relative gain array analysis). The simulations have also been used for testing various control structures and controllers. Controllers for the following types of control are suggested and tested:

Therefore only four, however essential, topologies are chosen for the comparison. The four topologies that have been selected are series, parallel, strigear and power split hybrid. The hybrid topologies have been simulated in two different cycles, the demanding highway cycle US06 and the slower urban cycle ECE15 with its much smoother accelerations.

The reference vehicle in the simulations has been a Toyota Prius, an electric hybrid family car, which is available on the market today. As input for the ICE, measured values from a SAAB naturally aspirated gasoline engine has been used, but scaled to better correspond to the ICE in the real Prius. There are many possible parameters in the simulation models, that are adjustable; vehicle chassis parameters, engine, electric machine(s) and battery size and types, losses models, charging strategies and driver behaviour etc. To investigate all of them is possibly interesting but not realistic in this survey. It is not the aim and the result flow would be overwhelming. Therefore six key parameters are chosen and thereafter adjusted one by one. The chosen ones are ICE dynamic response time constant, battery inner resistance, ICE charging gain, engine and motor sizes and finally maximum vehicle speed. The work presents a limited number of results. The results presented have been chosen to illustrate the impact the individual parameter has to the behaviour of the single topology.

The results of the simulated topologies have been compared with measurements made by MTC in Sweden and EPA in USA on a Toyota Prius.

The received results can be evaluated with the help of different criteria. Two different criteria are shown as an example of how the chosen criteria influence the results; the price of fuel consumption and produced emissions respectively a mutual comparison using weight factors.

The results from the simulations made, show that the parallel topology is the most efficient alternative. It is also the topology with the lowest complexity. It should be observed that the differences between the strigear and the parallel hybrid do sometimes not exist or turn to be of advantage to the strigear. The parallel topology however turns to be the preferred choice due to its lower complexity.

Automated Manufacture of Fertilizing Agglomerates from Burnt Wood Ash

Abstract :

In Sweden, extensive research is conducted to find alternative sources of energy that should partly replace the electric power production from nuclear power. With the ambition to create a sustainable system for producing energy, the use of renewable energy is expected to grow further and biofuels are expected to account for a significant part of this increase. However, when biofuels are burned or gasified, ash appears as a by-product. In order to overcome the problems related to deposition in land fills, the idea is to transform the ashes into a product - agglomerates - that easily could be recycled back to the forest grounds; as a fertilizer, or as a tool to reduce the acidification in the forest soil at the spreading area. This work considers the control of a transformation process, which transforms wood ash produced at a district heating plant into fertilizing agglomerates. A robust machine, built to comply with the industrial requirements for continuous operation, has been developed and is controlled by an industrial control system in order to enable an automated manufacture.

DC Distributed Power Systems - Analysis, Design and Control for a Renewable Energy System

Abstract :

Renewable energy systems are likely to become wide spread in the future due to environmental demands. As a consequence of the dispersed nature of renewable energy systems, this implies that there will be a distributed generation of electric power. Since most of the distributed electrical energy sources do not provide their electric power at line frequency and voltage, a DC bus is a useful common connection for several such sources. Due to the differences in output voltage among the sources, depending on both the type of source and their actual operating point, the sources are connected to the DC power system via power electronic converters. The intention behind the presented work is not to replace the existing AC power system, but to include local DC power systems. The AC and DC power systems are connected at some points in the network. The renewable energy sources are weak compared to the present hydro power and nuclear power plants, resulting in a need of power conditioning before the renewable energy is fed to the transmission lines. The benefit of such an approach is that power conditioning is applied on a central level, i.e. at the interface between the AC and DC power systems.

The thesis starts with an overview of related work. Present DC transmission systems are discussed and investigated in simulations. Then, different methods for load sharing and voltage control are discussed. Especially, the voltage droop control scheme is examined thoroughly. Since the droop control method does not require any high-speed communication between sources and loads, this is considered the most suitable for DC distributed power systems. The voltage feed back design of the controller also results in a specification of the DC bus capacitors (equivalents to DC link capacitors of single converters) needed for filtering. If the converters in the DC distribution system are equipped with capacitors selected from this design criterion and if the DC bus impedance is neglected, the source converters share the total load equally in per unit.

The same DC distribution bus configuration is studied in a wind power application. Especially the dynamic properties of load-source interactions are highlighted. They are interesting since the sources are considered weak for a distributed power system. This is illustrated with simulations where the power is fed from wind turbines only and constant power loads are controlled at the same time as the DC bus voltage level. The wind power generators are modeled as permanent-magnet synchronous machines. The controller needed for the machines, including position estimation and field weakening, is discussed. To control the DC bus voltage, the available wind power must be higher than the power consumed by the loads and the excess power removed by pitch angle control. Pitch angle control is a comparably slow process and, therefore, the DC bus voltage controller must handle the transient power distribution.

Personal safety and prevention of property damage are important factors of conventional AC power systems. For the investigated DC power system this is maybe even more important due to the fact that the star point of the sources and loads is left ungrounded or grounded through high impedance. The difficulty of detecting ground faults arises from the fact that the AC sources and loads are ungrounded or have high impedance to ground in order to effectively block zero-sequence currents flowing between the AC systems. A grounding scheme for the DC distribution system together with algorithms for detection of ground faults, are presented. The proposed method detects ground faults on both the AC and DC sides and is extended to cover short circuit faults with a minor work effort.

Two schemes for high voltage interconnection of DC systems are studied. One of them provides galvanic isolation, which is an advantage since elevated voltage might appear in the DC systems otherwise, in the case of a ground fault in the high voltage interconnection.

Experimental verifications follow the theoretical investigations introduced above. First, dynamic properties are studied and the behaviour predicted from theoretical analysis and simulations is verified. Then, load sharing is investigated. Also in this investigation, the experimental results agree with the simulated.

Dynamic Load Models for Power Systems - Estimation of Time-Varying Parameters During Normal Operation

Abstract :

Economic and environmental concerns will slow down the expansion of the transmission system in many countries. The addition of new transmission lines will be few and far between. The de-regulation of the power supply will introduce new power flow patterns on the bulk transmission systems. The net result is that the power systems will operate much closer to their transfer limits and operate there much longer time than has been necessary.

The risk for voltage collapse determines the transfer limits in many bulk transmission systems. The accurate determination of the transfer limits will be an increasingly important task to maintain the operational security and economic operation of the power system. Many studies have shown the importance of the load representation in voltage stability analysis. Static load models are not accurate enough for capturing the dynamics of the network. Therefore dynamic load models are needed even if voltage collapse, in many cases, is a slow phenomenon.

iii Due to the large amount of electric heating loads in Sweden and its effect on voltage stability, Hill and Karlsson have proposed a load model with exponential recovery. The model is expressed as a set of nonlinear differential equations, where the real and reactive load powers have a nonlinear dependency with voltage. The standard dynamic active load model is characterized by three parameters, steady state load-voltage dependence, transient load-voltage dependence and a load-recovery time constant. The same applies to reactive load. As an extension of the mentioned work, the present author proposes an automatic method for the determination of parameters in standard dynamic load models. The dynamic set of nonlinear equations has been linearised and the problem has been reduced to a linear identification problem. The Least Squares criterion is used for minimizing the error function between measured and simulated data.

Field measurements from continuous normal operation at the 20 kV and 50 kV-level from a substation in the South East of Sweden have provided over 1 GByte of data covering all seasons during the time period July 2001-June 2002. The determination of the load parameters based on this data has resulted in valuable information. The parameters’ time-varying characteristic and their dependency with weather and season of the year have been studied; there is correlation between the active and reactive recovery time constants, and between them and the corresponding steadystate characteristic of the load. Strong dependency of the transient active and reactive characteristic of the load with the temperature has been found.

Furthermore, some unexpected deviations in the reactive load parameters have led to a new representation of the reactive load. The reactive power level, which was previously used as normalization factor, is inappropriate. If instead apparent power level is used, the variability in the parameters that describe the reactive load response is drastically reduced.

Power System Oscillations - Detection, Estimation & Control

Abstract :

The topic of this thesis is the electro-mechanical oscillations which to some extent always are present in a power system. The demand of electric power is ever increasing. At the same time, the tolerance of disruptions in the power supply is decreasing. The deregulated market together with distributed generation have then pushed the system to operate during circumstances for which it was not designed. To this we can then add that getting concessions for new lines becomes more and more difficult in densly populated areas. All these factors makes the electric power system operate with smaller safety margins. Decreasing these margins limits with sustained availability is achieved by the application of advanced monitoring and control methods.

This thesis deals with this in several time-scales. When a large oscillation occurs it is important to detect it as fast as possible as the remedial action depends on if the current operating is due to a fault or an oscillation. In the thesis, a new method to distinguish these incidents from each other is presented.

In a slower time-scale it is important to monitor the dynamics of the electro-mechanical modes. This information can be used to verify that simulations correspond to the real world behaviour. Real-time methods can also be used to alarm operators or arm special protection schemes if the power system enters undesired operating conditions. A number of methods are studied and then evaluated on three case studies.

Finally, load modulation for damping enhancement is studied. It has previously been shown that modulation of active power at the transmission level increases the damping of the power system when correctly performed. Loads suitable for load modulation at the transmission level are rare and using actuators at the distribution level creates new problems. It is shown that it is possible to detect poor damping at the distribution level, thus reducing the need of communications. It is also shown how the variation of active and reactive power at the transmission level, caused by modulation of active power at the distribution level can be estimated without knowledge of the complete distribution network. This is most important as the active and reactive variations counteract each other. Finally load modulation at distribution level is evaluated on two test systems. It is shown that the damping is increased and that the influence of the reactive variation decreases the performance of load modulation. The degradation of the control scheme is, however, small in the studied cases.

Water Reuse in Paper Mills - Measurements and Control Problems in Biological Treatment

Abstract :

Paper manufacturing is a complex and multidisciplinary science due to the diversity of paper products, used raw materials and different production processes. Besides fibres different chemicals, water and energy are needed to produce paper. The use of fresh water has decreased significantly during the last decades and there are several reasons for this, such as: limited availability of fresh water, increased cost for effluent treatment and marketing benefits.

This decreased consumption has been made possible by the reuse of process water instead of fresh water. However, at a certain degree of closure different problems occur. Many of them are in some way related to the growth of microorganisms in the system. One method to solve the problems is to implement an internal kidney consisting of at least a biological treatment step. Since nutrients, such as nitrogen and phosphorous, normally are limited in the whitewater these have to be added in order to have an efficient biological treatment process. One major challenge is to operate the biological system with low concentrations of nutrients in the effluent otherwise the conditions in the whitewater system will be negatively affected. Consequently, there is a need for automatic control of the nutrient addition.

It is possible to control the flow of whitewater to the treatment process but not the actual concentrations of organic compounds in the whitewater, which therefore can be regarded as a process disturbance. An investigation was made at two different paper mills with different degrees of closure to determine the variation of chemical oxygen demand (COD) in the whitewater. The results showed that the whitewater concentration in an open mill could vary a lot whereas the conditions were more stable in a closed mill.

For the control there is a need for information about the process state and output from the system. In this case, for controlling a biological treatment of whitewater, different on-line instruments are needed. First of all, a market survey, limited to instruments for measurements of organic matter, ammonium and orthophosphate, was conducted. The experiences gathered about use of on-line instruments at several of the Swedish municipal treatments plants were explored in a telephone survey. One interesting observation was that most on-line instruments were only used for monitoring. The number of instruments used for direct control was low but this number was increasing as new and better instruments are becoming available. As a conclusion of these two surveys, three different brands of instruments were deemed suitable for measurements in whitewater.
Computer simulation is an important tool for evaluation of different controllers but requires a mathematical model of the system. Laboratory experiments were initiated to determine important parameters for such a model. Both mesophilic and thermophilic treatment of recycled fibre whitewater with a fluidised anaerobic reactor and an aerobic suspended biofilm process resulted in high removal of COD of around 90%. The nutrient requirement for the anaerobic mesophilic reactor was determined to 19 mg N/g CODreduced and 2.5 mg P/g CODreduced. For thermophilic degradation the requirement was determined to 24.5 mg N/g CODreduced and 4.4 mg P/g CODreduced for the anaerobic process and the corresponding values for the aerobic process were 37.1 mg N/g CODreduced and 5.5 mg P/g CODreduced. A decrease of the added amount of nitrogen to 77% of what was originally consumed did not have any immediate effect on the COD reduction.

Pilot tests with the purpose to study both the stability of a biological treatment process and evaluate two different on-line instruments were conducted at a packaging board mill. The results demonstrated that the removal efficiency was not markedly affected from variations of the load to the combined anaerobic/aerobic treatment process and that both instruments failed to provide stable results. Experiences from other instruments have been gathered during the assembly of a complete system consisting of a pilot plant of a biological treatment process, on-line instruments and data-acquisition equipment.
It has been demonstrated that it is possible to use on-line instruments for measurements in whitewater to acquire information about the biological treatment process. This information could be used in several different ways for the control of the addition of nutrients. Different control structures are suggested ranging from feed forward of the organic load with corrective feedback of concentrations in the anaerobic effluent to more complex model-based control structures with automatic update of model parameters.

Improved Models for DC-DC Converters

Abstract :

To obtain high performance control of a dc-dc converter, a good model of the converter is needed. It is suitable to consider the load to be included in the converter model since it usually affects the dynamics. The load is often the most variable part of this system. If the load current and the output voltage are measured there are good possibilities to obtain a good model of the load on-line. Adaptive control can then be applied to improve the control.
 
In peak current-mode control, the output voltage and the inductor current are measured and utilized by the controller. This thesis analyses some properties that can be obtained if the load current is also measured and utilized for control. Accurate expressions for the control-to-output transfer function, the output impedance, and the audio susceptibility are derived for the buck, boost, and buck-boost converters operated in continuous conduction mode in the case where the load is a linear resistor. If the measured load current is utilized properly by the controller, the output impedance becomes low and the control-to-output transfer function becomes almost invariant for different loads. The use of load current acts as a feedforward term if the load is a current source. However, if the load is a resistor the load current is influenced by changes in the output voltage and the stability is affected. Therefore, the use of load current is not a feedforward action in this case. Instead it can be seen as gain scheduling, which can be considered a special case of adaptive control.
 
In the thesis it is also shown that the two published models for current-mode control, Ridley (1991) and Tan and Middlebrook (1995), give accurate expressions for the control-to-output transfer function and the output impedance but not for the audio susceptibility. A novel model for the audio susceptibility is presented and it is used to improve the two published models.
 
Most of the results in the thesis are validated by comparing the frequency responses predicted by the expressions and switched large-signal simulation models.

DC-DC Converters - Dynamic Model Design and Experimental Verification

Abstract :

To obtain high performance control of a dc-dc converter, a good model of the converter is needed. The load usually affects the dynamics and one way to take this into consideration is to regard the load as a part of the converter. The load is often the most variable part of this system. If the load current and the output voltage are measured there are good possibilities to obtain a good model of the load on-line. Adaptive control can then be applied to improve the control. In peak current-mode control, the output voltage and the inductor current are measured and utilized for control. In the author’s licentiate thesis, analytic models were derived for the case where the load current is also measured and utilized for control.

The control-to-output transfer function, the output impedance, and the audio susceptibility were derived for the buck, boost, and buck-boost converters operated in continuous conduction mode in the case of resistive load. The use of load current can be seen as gain scheduling in the case where the load is a resistor. Gain scheduling can be considered a special case of adaptive control. The majority of the results in the licentiate thesis were validated by comparing the frequency responses predicted by the analytic models and switched large-signal simulation models.

In this thesis, additional results are presented for the buck converter. Experimental results obtained by means of a network analyzer verify the derived control-to-output transfer function and the audio susceptibility but not the output impedance at low frequencies. In the experimental buck converter there are stray resistances in the inductor, transistor, and diode but these stray resistances were not considered in the licentiate thesis. A new transfer function for the output impedance is derived where these stray resistances are considered and it is in good agreement with the experimental result also at low frequencies.

If the current to the output capacitor is measured in addition to the output voltage and the inductor current, the load current can be calculated as the difference between the inductor and capacitor currents in the case of the buck converter. Hence, the measurement of the load current can be replaced by measurement of the capacitor current. If this possibility is utilized and the capacitor current is measured by means of a current transformer, a low-frequency resonance is introduced in the frequency responses according to experimental results. The reason for this resonance is due to the high-pass-filter characteristics of the current transformer. A new analytic model is derived which predicts the resonance.

Design, Modelling and Control of Electrical Machines - With Applications to Iron-powder Machines and Acoustic Noise

Abstract :

This thesis consists of two parts, the first dealing with the design of iron-powder synchronous machines, and the second with the analysis and prediction of the acoustic noise in electrical machines.

In Part I, a 1.6 kW electrically magnetized claw-pole machine with magnetically conducting end-plates has been analyzed and a prototype tested. The machine is built from soft magnetic composite material (SMC), also known as iron-powder. The magnetic isotropy of SMC gives enormous flexibility in electrical machine design, enabling new topologies exploiting three dimensional flux paths. This is the main advantage compared to conventional machines using laminations, where the flux is constrained into two dimensions. The novelty of the machine presented lies in that the slip-rings in the rotor are no longer needed, since the field coils are removed from the rotor and placed in magnetically conducting end-plates attached to both sides of the stator. This also improves the cooling capability of the copper losses from the field winding, allowing an increased electric loading. The rotor is of the claw-pole type, and the end-plates close the magnetic circuit between the stator and the rotor. The machine has been optimized using a magnetic equivalent circuit model allowing rotation, where non-linearities have been included using an iterative approach based on the linearisation of the BH curve. The traditional leakage paths in claw-pole machines are modified because of the magnetically conducting end-plates, and alternatives are proposed to reduce them. The machine has also been compared to two alternative topologies with electrical magnetization and another with permanent magnets. The comparison has been carried out for a similar temperature rise in the windings, and thermal models have been developed for every machine to determine their maximum electric loading. The rotational and alternating components of the iron losses are calculated using the finite element method (FEM). The results from the measurements indicate that the average torque is 14% lower than predicted. This is probably due to a leakage path between the end-plates through the shaft, which carries the homopolar flux, and that was not considered in the predictions. The concept of series magnetization has also been tested. This consists of feeding the field winding directly from the rectified three phase armature currents at the neutral point, therefore eliminating the need for the d.c. power source.

Electrical machines are finding application in new environments where lower noise levels are demanded. The main focus of Part II is the measurement and prediction of the noise emissions from induction motors using the vector control technique as well as the analysis of some structural changes to reduce these emissions. A digital drive system has been developed for a 2.2 kW induction motor, and its dynamic capabilities demonstrated for a wide range of the frequency spectrum. This tool has been used for the experimental evaluation of the noise emissions when the flux and/or the torque are modulated with high frequency noise signals. The results showed that the noise emissions were higher when the flux was modulated compared to the torque, although the differences were considerably reduced when the machine was loaded. It was also observed that the noise emissions were decreased importantly at load. Sound pressure and sound intensity measurements have been conducted with the rotor stationary and rotating at low speed, showing that the most proper way to quantify the noise emissions from electric machinery is to measure the sound power. A method for the prediction of the noise emissions has been proposed, based on the interactive use of commercial packages for mechanical, electromagnetic and acoustic analyses based in the finite and boundary element methods. The results show that the accuracy of the noise prediction depends on the proper calculation of the modes of vibration in the structural analysis, as well as a suitable selection of the material damping. Skewing also needs to be modelled in order to account for the rotor harmonics. A study has been conducted to assess the effectiveness of introducing peripheral air gap layers around the stator core to reduce the noise emissions. It was observed that the acoustic behaviour was not improved, since the reduction of the stiffness in the outer part of the core actually increased its sensitivity to the vibrations.

Design of Powder Core Motors

Abstract :

The goal of the study presented in this thesis is to evaluate the advantages and drawbacks of using powder technology in the design of the iron core of small claw-pole electric motors. The use of soft magnetic composites (SMC) and compaction technology allows the creation of complex 3D iron cores. The additional dimension opens for new solutions of the electromechanical energy conversion. A claw-pole motor among the transversal flux machines that has particularly high specific torque is in the focus of research interest. Generally, as the iron core can be more complicated, the winding is chosen to be simpler in the powder core motors. The thesis focuses on the machine design of a single-phase and a two-phase low-power claw-pole motor. The predicted results compare well with measurements of the prototype motors.

The motor design process in this thesis uses a magnetic equivalent circuit (MEC) model of the outer-rotor claw-pole motors that is accurate enough to describe the physics of the electromagnetic conversion. Additional equivalent circuits are made to evaluate the mechanic and thermal loading of the machines. The outcome of the equivalent circuit models is enough to estimate roughly the optimal size of the motor and the motor output according to the materials selected.

After the rough design process, which is based on equivalent circuits, is finished, a series of FE magnetostatic analyses are made in order to evaluate the static characteristics of the motors, to specify the magnetization losses and to carry out a sensitivity study for the proposed size of the motors. Finally, the magnetic, mechanic and thermal design is analyzed dynamically and statically by the use of coupled multiphysics. The task of the coupled multiphysics is to find out the cooling capability and the thermal limit of the motor as well as the mechanic stress in the motor parts due to magneto-mechanic loading. It is discussed how the discrepancy between the calculated and measured cogging torque depends on the fineness of the 3D FE air gap mesh.

Iron loss estimation based on the results of the FE-analysis is made taking the local rotation, and not only pulsation, of the magnetic flux into consideration. It is shown that the loss coefficients in the material model must be adapted to account for flux rotation.

A part from the output of the machine as an electromechanical energy converter is their controllability in the electric drive system. Based on the static characteristics, which are calculated in the FE-analysis and verified in prototype measurements, a tailor made control method is developed for the machines designed. Results are presented of extensive simulations and experimental verifications of the proposed control strategy and power electronic circuitry. The high-speed four-pole single-phase motor shows satisfactory results. The other motor, which has 20 poles and two phases, has a main weakness in its complex assembling and a large cogging torque.

Dynamic Power System Load - Estimation of Parameters from Operational Data

Abstract :

The significance of load modeling for voltage stability studies has been emphasized by several disturbances, which have taken place in the past years. They have shown that the loads in combination with other dynamics are among the main contributors of prolonged low voltage conditions, voltage instability and collapse in the power system. As a result of these disturbances new investigations have come up to better understand the nature of the load. However, power system loads keep being very difficult to model; the load generally aggregates a large number of individual components of different nature, different load dynamics are excited depending on the time frame of actuation and the type of disturbance affecting the system, and the load is highly dependent on external factors such as weather conditions.

This thesis investigates the load-voltage characteristic during two different time scales, long-term over several minutes, and short-term covering ms to several seconds, for different sized disturbances, and its impact on the calculation of transfer limits and security margins in voltage stability studies. The accurate determination of transfer limits will be an increasingly important task to maintain the operational security and economic dispatch of the power system. The location of the stability limits and the determination of transfer limits depend on the load-voltage characteristic since load relief due to the load-voltage dependency results in larger transfer limits. Moreover, the importance of using dynamic load models instead of static ones in stability studies is highlighted in this thesis.

Due to the large amount of electrical heating loads in Sweden and its effect on voltage stability, a dynamic load model with exponential recovery, previously proposed by Hill and Karlsson, [Karlsson and Hill, 1994], has been the starting point for the investigations. Field measurements from continuous normal operation at the 20 kV-level from a substation in Sweden have provided a large amount of data covering all seasons during the time period July 2001-June 2002, and have resulted in extensive, unique and interesting recordings of active and reactive load characteristic and its dependency with small voltage variations. The data have revealed the variation of the load parameters and their dependency with weather and season of the year. The work has also contributed to a better approach for the normalization of traditional reactive load models.

Furthermore the load-voltage characteristic during large disturbances has been investigated based on field measurements of phase-to-phase faults in a non-effectively earthed 50 kV system in Sweden. Three-phase currents and voltages have been used to estimate the active and reactive power. The recordings exhibited voltage dips up to 30% in the positive sequence voltage. The severity of the disturbances accentuates the nonlinear behavior of the load; the active and reactive power rapidly increase after fault clearing to levels even above the pre-disturbance value due to the reacceleration of motors. The full recovery of the voltage is delayed due to the re-connection of tripped load. Moreover, it is shown that traditional load models do not accurately reflect the load behavior during these disturbances, for voltage dips around 12 % or larger due to the nonlinearities. An alternative load model, which represents the nonlinearities, has been tested. The superior behavior is demonstrated with the field measurements.

Finally, some guidelines for industry to better account for the load in future stability studies have been included as a corollary of this thesis.

Control of Hybrid Electric Vehicles with Diesel Engines

Abstract :

This thesis is an approach to improve electric hybrid vehicles with respect to fuel consumption and to fulfil the future intended NOx emission regulations. It is based upon the conclusions made in the licentiate thesis Analysing Hybrid Drive System Topologies (Jonasson, 2002).

The study in this thesis is restricted to a parallel hybrid vehicle equipped with a diesel engine, two electric machines and electrical energy storage and a model thereof is presented in the thesis.

The choice to focus on the diesel engine is related to the high efficiency of this engine that also is the reason for the in later years increased market for diesel engines in conventional vehicles. Since one of the disadvantages, related to the diesel engine, are the nitrogen oxides (NOx) emissions, efforts is concentrated on reducing them, by means of the advantages of hybridisation.

The reference vehicle in the simulations presented in this thesis is a Toyota Prius, an electric hybrid passenger car, which is available on the market today. As input for the combustion engine model, engine data from a diesel engine considered as state of the art 2004, has been used. The engine data is scaled to correspond to the engine size used in the Prius. It should be mentioned that the engine in the Toyota Prius is run on petrol.

There are many possible parameters in the simulation model, which are adjustable; vehicle chassis parameters, engine, electric machine(s) and battery size and types, losses models, charging strategies and driver behaviour etc. A number of key parameters have been selected in this study: control strategy, NOx control by means of EGR (exhaust gas recirculation) and SCR (selective catalytic reduction), gear ratios and gearshift strategies and finally cylinder deactivation. The accuracy of the simulation model is ratified by means of measured data on the engine used in the simulation.

Fuel consumption and NOx are determined by using look-up-tables based on measured data. The engine temperature, needed to determine the NOx conversion by means of SCR, is also received from a look-up-table.

The simulation model is evaluated in the driving cycle ECE+EUDC.

The results presented are chosen to illustrate the impact each individual parameter has on the behaviour of the hybrid vehicle, the fuel consumption and the emissions.

The results from the simulations show that it is possible to pass the expected limit of the future Euro 5 NOx regulations, if NOx emission treatment with EGR and SCR is implemented. The price to pay for this action is to sacrifice some of the fuel savings that the hybridization brings. The result is nevertheless a vehicle with decreased fuel consumption compared with a conventional diesel powered vehicle, and a vehicle that passes the intended emission regulation.

Failure Monitoring and Asset Condition Assessment in Water Supply Systems

Abstract :
In this thesis, different aspects of failure management in urban water supply
systems are discussed. As assets are getting older, the number of pipe
failures is increasing. Therefore, an efficient failure management strategy
becomes important. Two types of failure management strategies can be applied:
proactive asset condition assessment to prevent a failure and reactive
failure detection and location to minimize the reaction time and losses associated
with a failure. Currently available condition assessment techniques
cannot be extensively applied in water supply systems due to high cost and
slow speed of inspection. Existing failure detection and location approaches
do not allow for quick reaction to failures. Thus, there is a need for systematic
inspection and monitoring techniques. Automatic failure monitoring and
systematic asset condition assessment methods are presented in this thesis.
Due to the different topology and hydraulic characteristics of the transmission
(pipelines) and distribution (network) components of a water supply
system, separate failure detection and location techniques are proposed. For
both pipeline and network cases, two types of failures are considered - sudden
pipe ruptures and breaks that develop over a longer period of time. For
the pipeline case, a periodical leak diagnosis system based on transient response
difference monitoring, is presented together with a burst monitoring,
detection and location system, which is designed for rapid reaction to sudden
pipeline ruptures. A single continuous pressure monitoring station is
sufficient to detect and locate a failure along the whole length of a pipeline.
Proposed systems have been validated in laboratory and field conditions.
For the network case, two continuous failure monitoring approaches are developed,
which are based on steady-state (first method) and unsteady-state

Differential Protection for Arbitrary Three-Phase Power Transformers

Abstract :
This thesis describes how to provide standardized, current based, differential protection for any three-phase power transformer, including phase-shifting transformers with variable phase angle shift and transformers of all construction types and internal on-load tapchanger configurations. The use of standard transformer differential protection for such applications is considered impossible in the protective relaying standards and practices currently applied. The first part of the thesis provides the background for different types of power transformers and the differential protection schemes currently applied. After that a complete mathematical proof for the universal transformer differential protection principle, based on theory of symmetrical components, is derived. It is demonstrated that it is possible to make numerical differential protection relays which can properly calculate differential currents for any power transformer, regardless of whether it is of fixed or variable phase angle shift construction and whether current magnitude variations are caused by on-load tap-changer(s). It is shown how to correctly calculate differential currents by simultaneously providing on-line compensation for current magnitude variations, on-line compensation for arbitrary phase angle shift variations and settable zero-sequence current reduction on any power transformer side. By using this method differential protection for arbitrary power transformers will be ideally balanced for all symmetrical and non-symmetrical through-load conditions and external faults. The method is independent of individual transformer winding connection details (i.e. star, delta or zigzag), but dependent on having the correct information about actual on-load tap-changer(s) position if they are built-in within the protected power transformer. The implementation and practical use of this new universal principle is quite simple, as all necessary transformer data is commonly available on the protected power transformer rating plate. Practical application of the universal method for the differential protection of standard transformers, special transformers and phase shifting transformer is presented. Detailed testing of this new universal differential protection method is given and it is based on actual field recordings captured by numerical relays in existing phase-shifting transformer installations and on simulations from the Real Time Digital Simulator for a practical dual-core, symmetrical phase-shifting transformer. The implementation of the universal transformer differential method for analogue and numerical transformer differential relays is also described. Problems for the differential protection caused by transformer inrush currents are discussed. The mathematical relationship between differential protection and directional protection is derived. Then it is shown that through the addition of supplementary directional criteria, security and speed of the operation of the transformer differential protection can be improved. Finally, the use of additional directional criteria to significantly improve the sensitivity of the differential protection for transformer winding turn-to-turn faults is suggested. Captured disturbance files from numerical differential relays in actual power transformer installations, during internal and external faults, have been used to demonstrate the performance of additional directional criteria.