Active magnetic bearings (AMBs) have become a key technology in industrial applications with a continued drive for cost reduction and an increase in reliability. AMBs require position feedback to suspend the rotor. Conventional contactless position sensors are used to measure the rotor's position. The major disadvantages of conventional position sensors are their cost and that the sensors are viewed as a weak point in an AMB system. A self-sensing sensor is a type of sensor which is cost effective, reduces sensor wire-length and increases reliability, thus ideal for the industry. This type of sensor relies on the current and voltage signals of the AMB's to obtain the rotor position. Due to the rapid and advanced development of digital electronics, it has become more powerful and cheaper, thus self-sensing in mass production will be cost effective. Different self-sensing approaches were developed in the past and can be divided into two main categories: state estimation and amplitude modulation approaches. In this research the focus will be on the amplitude modulation approach. Amplitude modulation makes use of two signals, namely the modulation signal and the carrier signal. In a self-sensing AMB system the carrier can be a high frequency component injected into the system or the switching ripple of the switch mode power amplifier can be used. The modulation signal is the change in rotor position which results in changing inductances. The actuator material introduces nonlinear effects on the estimated position. Due to these nonlinear effects, it is rather difficult to obtain the rotor position. The first industrial application of a self-sensing turbomolecular pump system was implemented in 2005 by S2M. The aim of this thesis is to evaluate existing self-sensing schemes, devise improvements and investigate possible new schemes. Four different demodulation methods and two new self-sensing schemes are evaluated. An AMB transient simulation model which includes saturation, hysteresis, eddy currents and cross-coupling is used to evaluate the schemes in simulation. The self-sensing schemes are implemented in hardware and evaluated on a 7 A rms 500 N AMB. A comparative study was done on the different self-sensing schemes. From the comparative study it was determined that the gain- and phase effects have a direct effect on the sensitivity of the system. It was also proved that self-sensing can be implemented on a coupled AMB with a sensitivity of 10.3 dB. / Thesis (Ph.D. (Electrical and Electronic Engineering))--North-West University, Potchefstroom Campus, 2009.

The study presents a pragmatic evaluation of the influence of inductive teaching on grade 11 learners in two high schools in Tshwane West District in the Gauteng province in a form of pseudo experiment complemented with a qualitative investigation. The study focussed on the influence of inductive teaching on the nature of conceptualisation of and the learning achievement with regard to functions in Grade 11. A model adopted by O'Callaghan that identifies and applies the four competencies of modelling a function, interpreting a function, translating and reifying a function proved to be relevant for the investigation and hence was adapted for the study. The methodology used included data collection through pretest-posttest control group experimental design complemented with unstructured interviews. The verification of the reliability of research instruments and data analysis was done with the assistance of the Northwest University (Potchefstroom Campus) Statistical Consultation Services and through identification of common perceptions and experiences of participants. The results of the study did indicate positive influence of inductive teaching on the nature and quality of conceptual learning of the function concept. / Thesis (M.Ed.)--North-West University, Potchefstroom Campus, 2009.

Active magnetic bearings (AMBs) have become a key technology in industrial applications with a continued drive for cost reduction and an increase in reliability. AMBs require position feedback to suspend the rotor. Conventional contactless position sensors are used to measure the rotor's position. The major disadvantages of conventional position sensors are their cost and that the sensors are viewed as a weak point in an AMB system. A self-sensing sensor is a type of sensor which is cost effective, reduces sensor wire-length and increases reliability, thus ideal for the industry. This type of sensor relies on the current and voltage signals of the AMB's to obtain the rotor position. Due to the rapid and advanced development of digital electronics, it has become more powerful and cheaper, thus self-sensing in mass production will be cost effective. Different self-sensing approaches were developed in the past and can be divided into two main categories: state estimation and amplitude modulation approaches. In this research the focus will be on the amplitude modulation approach. Amplitude modulation makes use of two signals, namely the modulation signal and the carrier signal. In a self-sensing AMB system the carrier can be a high frequency component injected into the system or the switching ripple of the switch mode power amplifier can be used. The modulation signal is the change in rotor position which results in changing inductances. The actuator material introduces nonlinear effects on the estimated position. Due to these nonlinear effects, it is rather difficult to obtain the rotor position. The first industrial application of a self-sensing turbomolecular pump system was implemented in 2005 by S2M. The aim of this thesis is to evaluate existing self-sensing schemes, devise improvements and investigate possible new schemes. Four different demodulation methods and two new self-sensing schemes are evaluated. An AMB transient simulation model which includes saturation, hysteresis, eddy currents and cross-coupling is used to evaluate the schemes in simulation. The self-sensing schemes are implemented in hardware and evaluated on a 7 A rms 500 N AMB. A comparative study was done on the different self-sensing schemes. From the comparative study it was determined that the gain- and phase effects have a direct effect on the sensitivity of the system. It was also proved that self-sensing can be implemented on a coupled AMB with a sensitivity of 10.3 dB. / Thesis (Ph.D. (Electrical and Electronic Engineering))--North-West University, Potchefstroom Campus, 2009.

The study presents a pragmatic evaluation of the influence of inductive teaching on grade 11 learners in two high schools in Tshwane West District in the Gauteng province in a form of pseudo experiment complemented with a qualitative investigation. The study focussed on the influence of inductive teaching on the nature of conceptualisation of and the learning achievement with regard to functions in Grade 11. A model adopted by O'Callaghan that identifies and applies the four competencies of modelling a function, interpreting a function, translating and reifying a function proved to be relevant for the investigation and hence was adapted for the study. The methodology used included data collection through pretest-posttest control group experimental design complemented with unstructured interviews. The verification of the reliability of research instruments and data analysis was done with the assistance of the Northwest University (Potchefstroom Campus) Statistical Consultation Services and through identification of common perceptions and experiences of participants. The results of the study did indicate positive influence of inductive teaching on the nature and quality of conceptual learning of the function concept. / Thesis (M.Ed.)--North-West University, Potchefstroom Campus, 2009.

Crosstalk is defined as an unwanted coupling between the conductors. By this it is meant that signals from one of the signal conductors (a generator in this case) are coupled to another signal conductor (receptor), or conductors (receptors), depending on the number of conductors in the vicinity of the generator. Crosstalk in this way affects the signal level on the receptor and thereby affects the total system performance within the system. This can happen in several ways, one of which is through edge coupling. Edge coupling is a process where two signal conductors are placed beside each other in the same layer while the ground conductor could have been placed either under these conductors, in a separate layer like Mclin (Microstrip coupled lines) and Sclin (Coupled striplines), or beside the signal conductors as in Cpwcpl2 (Coplanar wave guide coupled lines). This then means that edge coupling occurs through the sides where the generator and the receptor are facing each other. Broadside coupling is another way, where it occurs when the signal conductors are broadside faced to each other in different layers with reference planes above and below these signal conductors. Coupling of the signals from the generator to the receptor can occur through capacitive coupling or inductive coupling. Capacitive coupling, also known as electrical coupling, occurs due to the difference in the characteristic impedance of the generator (usually 50 or 100 §Ù) and its heavy load (1 k§Ù or more) which results in high voltage difference between the generator and the reference conductor (ground). This leads to the creation of a charge across the generator and the receptor-facing sides and finally results in the electric field coupling between them. On the other hand, inductive coupling, also known as magnetic coupling, occurs when the load is less than the characteristic impedance of the generator, and this thereby results in a heavy current flow through the generator which in return results in a strong magnetic field around itself and so leads to magnetic coupling to the receptor. The aim in this thesis is to measure both the capacitive and inductive coupling load¡¯s impacts on both the edge coupling and the broadside coupling models through crosstalk on the receptor. This thesis starts with the background and corresponding theory and equations to the crosstalk coupling. Later on it tests both the edge- and broadside coupling models with different physical properties exploitation. Inductive and capacitive loads are used to measure the resulting crosstalk coupling. Particularly to see the effect of capacitive and inductive coupling in reality in multi layered PCB, a Sbclin (Broadside coupled striplines) model has been used with different angular placement of the generator. Finally mclin physical models are compared with the simulated models and corresponding differences are discussed. It can be concluded that crosstalk effect increases or decreases with physical properties exploitation. Crosstalk also increases with the wrong termination of the load.

Algorithms that learn sets of rules describing a concept from its examples have been widely studied in machine learning and have been applied to problems in medicine, molecular biology, planning and linguistics. Many of these algorithms used a separate-and-conquer strategy, repeatedly searching for rules that explain different parts of the example set. When examples are scarce, however, it is difficult for these algorithms to evaluate the relative quality of two or more rules which fit the examples equally well. This dissertation proposes, implements and examines a general technique for modifying rule evaluation in order to improve learning performance in these situations. This approach, called Description-based Evaluation Function Transfer (DEFT), adjusts the way rules are evaluated on a target concept by taking into account the performance of similar rules on a related support task that is supplied by a domain expert. Central to this approach is a novel theory of task similarity that is defined in terms of syntactic properties of rules, called descriptions, which define what it means for rules to be similar. Each description is associated with a prior distribution over classification probabilities derived from the support examples and a rule's evaluation on a target task is combined with the relevant prior using Bayes' rule. Given some natural conditions regarding the similarity of the target and support task, it is shown that modifying rule evaluation in this way is guaranteed to improve estimates of the true classification probabilities. Algorithms to efficiently implement Deft are described, analysed and used to measure the effect these improvements have on the quality of induced theories. Empirical studies of this implementation were carried out on two artificial and two real-world domains. The results show that the inductive transfer of evaluation bias based on rule similarity is an effective and practical way to improve learning when training examples are limited.

Real-world learning tasks present a range of issues for learning systems. Learning tasks can be complex and the training data noisy. When operating as part of a larger system, there may be limitations on available memory and computational resources. Learners may also be required to provide results from a stream. This thesis investigates the problem of incremental, relational learning from imperfect data with constrained time and memory resources. The learning process involves incremental update of a theory when an example is presented that contradicts the theory. Contradictions occur if there is an incorrect theory or noisy data. The learner cannot discriminate between the two possibilities, so both are considered and the better possibility used. Additionally, all changes to the theory must have support from multiple examples. These two principles allow learning from imperfect data. The Minimum Description Length principle is used for selection between possible worlds and determining appropriate levels of additional justification. A new encoding scheme allows the use of MDL within the framework of Inductive Logic Programming. Examples must be stored to provide additional justification for revisions without violating resource requirements. A new algorithm determines when to discard examples, minimising total usage while ensuring sufficient storage for justifications. Searching for revisions is the most computationally expensive part of the process, yet not all searches are successful. Another new algorithm uses a notion of theory stability as a guide to occasionally disallow entire searches to reduce overall time. The approach has been implemented as a learner called NILE. Empirical tests include two challenging domains where this type of learner acts as one component of a larger task. The first of these involves recognition of behavior activation conditions in another agent as part of an opponent modeling task. The second, more challenging task is learning to identify objects in visual images by recognising relationships between image features. These experiments highlight NILE'S strengths and limitations as well as providing new n domains for future work in ILP.

Η παρούσα διπλωματική εργασία πραγματεύεται τη μελέτη μιας διάταξης ανεμογεννήτριας με επαγωγική μηχανή βραχυκυκλωμένου κλωβού η οποία συνδέεται μέσω dc διασύνδεσης με το δίκτυο. Τη dc διασύνδεση τη δημιουργεί ένας dc πυκνωτής και από την πλευρά της μηχανής υπάρχει ένας ac/dc πλήρως ελεγχόμενος ανορθωτής ισχύος ενώ από την πλευρά του δικτύου βρίσκεται ένας dc/ac πλήρως ελεγχόμενος αντιστροφέας ισχύος. Σκοπός της εργασίας είναι αρχικά να μελετήσουμε θεωρητικά το συνολικό σύστημα της ανεμογεννήτριας χρησιμοποιώντας το μετασχηματισμό Park για τη μοντελοποίηση του στο d-q πλαίσιο αναφοράς. Στη συνέχεια προχωρούμε στον υπολογισμό των εξισώσεων στο χώρο κατάστασης και τέλικά προτείνουμε ένα σύστημα ελέγχου το οποίο βασίζεται στους PI ελεγκτές. Τέλος, προσομοιώνουμε το παραπάνω σύστημα με τη βοήθεια της εφαρμογής Simulink του λογισμικού Matlab και μελετώντας τα αποτελέσματα, εξάγουμε τα ανάλογα συμπεράσματα. / The present thesis deals with the study of provision of wind generator with inductive machine of which is connected via dc interconnection with the network. A dc capacitor creates the dc interconnection with the network.The aim of this work is initially to study theoretically the total system of wind generator using the transformation Park for its modelling in the d-q frame of report. Afterwards we advance in the calculation of equations in the space of situation and finally we propose a system of control which is based on PI controllers. Finally, we simulate the system with the help of the Simulink application of Matlab and studying the results, we export the proportional conclusions.

This thesis analyzes the framing of the economic policies of American populists Donald Trump and Bernie Sanders in American quality newspapers during the 2016 primary elections. This is done with the intent of investigating if the common assumption that the economic policies of pop-ulists are irresponsible could be observed in the news media. By analyzing two newspapers oppo-site each other on the political spectrum – The New York Times to the left and The Wall Street Journal to the right – and their coverage of the policies of two candidates opposite each other on the political spectrum, a complex comparative study is achieved. An inductive frame analysis yielded six different frames used when discussing the economic policies of the candidates: dam-aging, ignorant, unpredictable, unorthodox, status quo and positive change. While none of the frames fit into a binary “irresponsible” or “not irresponsible” category, the frequent use of the damaging and ignorant frames suggests that the assumption of populist irresponsibility is alive and well.

It is well acknowledged in the AI community that finding a good reward function for reinforcement learning is extremely challenging. Apprenticeship learning (AL) is a class of “learning from demonstration” techniques where the reward function of a Markov Decision Process (MDP) is unknown to the learning agent and the agent uses inverse reinforcement learning (IRL) methods to recover expert policy from a set of expert demonstrations. However, as the agent learns exclusively from observations, given a constraint on the probability of the agent running into unwanted situations, there is no verification, nor guarantee, for the learnt policy on the satisfaction of the restriction. In this dissertation, we study the problem of how to guide AL to learn a policy that is inherently safe while still meeting its learning objective. By combining formal methods with imitation learning, a Counterexample-Guided Apprenticeship Learning algorithm is proposed. We consider a setting where the unknown reward function is assumed to be a linear combination of a set of state features, and the safety property is specified in Probabilistic Computation Tree Logic (PCTL). By embedding probabilistic model checking inside AL, we propose a novel counterexample-guided approach that can ensure both safety and performance of the learnt policy. This algorithm guarantees that given some formal safety specification defined by probabilistic temporal logic, the learnt policy shall satisfy this specification. We demonstrate the effectiveness of our approach on several challenging AL scenarios where safety is essential.