The development of an online method for the early detection of a stator turn insulation deterioration is the objective of the research at hand. A high percentage of motor breakdowns is related to the failure of the stator insulation system. Since most of the stator insulation failures originate in the breakdown of the turn-to-turn insulation, the research in this realm is of great significance. Despite the progress that has been made in the field of stator turn fault detection methods, the most popular and the best known ones are still limited to the detection of solid turn faults. The time span between a solid turn fault and the breakdown of the primary insulation system can be as short as a few seconds. Therefore, it is desirable to develop a method capable of detecting the deterioration of the turn insulation as early as possible and prior to the development of a solid turn fault. The different stresses that cause the aging of the insulation and eventually lead to failure are described as well as the various patterns of an insulation failure. A comprehensive literature survey shows the methods presently used for the monitoring of the turn insulation. Up to now no well-tested and reliable online method that can find the deterioration of the turn insulation is available. The most commonly used turn insulation test is the surge test, which, however, is performed only when the motor is out of service and disconnected from the supply. So far no research at all has been conducted on the application of an online surge test. The research at hand examines the applicability of the surge test to an operating machine. Various topologies of online surge testing are examined with regard to their practicability and their limitations. The most practical configuration is chosen for further analysis, implementation and development. Moreover, practical challenges are presented by the non-idealities of the induction machine like the eccentricity of the rotor and the rotor slotting, and have to be taken into account. Two solutions to eliminate the influence of the rotor position on the surge waveform are presented. Even though the basic concepts of online surge testing can be validated experimentally by a machine with a solid turn fault, it is preferable to use a machine with a deteriorated turn insulation. Therefore, a method, which does not require complex and expensive hardware, to experimentally emulate the turn insulation breakdown is implemented. The concepts at any stage of the work are supported by simulations and experimental results. In addition, the theory of surge testing is further developed by giving new definitions of the test's sensitivity, i.e., the frequency sensitivity and the error area ratio (EAR) sensitivity.

In industry, it is crucial to be able to detect damage or abnormal behavior in machines. A machine's downtime can be minimized by detecting and repairing faulty components of the machine as early as possible. It is, however, economically inefficient and labor-intensive to detect machine fault sounds manual. In comparison with manual machine failure detection, automatic failure detection systems can reduce operating and personnel costs.  Although prior research has identified many methods to detect failures in drill machines using vibration or sound signals, this field still remains many challenges. Most previous research using machine learning techniques has been based on features that are extracted manually from the raw sound signals and classified using conventional classifiers (SVM, Gaussian mixture model, etc.). However, manual extraction and selection of features may be tedious for researchers, and their choices may be biased because it is difficult to identify which features are good and contain an essential description of sounds for classification. Recent studies have used LSTM, end-to-end 1D CNN, and 2D CNN as classifiers for classification, but these have limited accuracy for machine failure detection. Besides, machine failure occurs very rarely in the data. Moreover, the sounds in the real-world dataset have complex waveforms and usually are a combination of noise and sound presented at the same time. Given that drill failure detection is essential to apply in the industry to detect failures in machines, I felt compelled to propose a system that can detect anomalies in the drill machine effectively, especially for a small dataset. This thesis proposed modern artificial intelligence methods for the detection of drill failures using drill sounds provided by Valmet AB. Instead of using raw sound signals, the image representations of sound signals (Mel spectrograms and log-Mel spectrograms) were used as the input of my proposed models. For feature extraction, I proposed using deep learning 2-D convolutional neural networks (2D-CNN) to extract features from image representations of sound signals. To classify three classes in the dataset from Valmet AB (anomalous sounds, normal sounds, and irrelevant sounds), I proposed either using conventional machine learning classifiers (KNN, SVM, and linear discriminant) or a recurrent neural network (long short-term memory). For using conventional machine learning methods as classifiers, pre-trained VGG19 was used to extract features and neighborhood component analysis (NCA) as the feature selection. For using long short-term memory (LSTM), a small 2D-CNN was proposed to extract features and used an attention layer after LSTM to focus on the anomaly of the sound when the drill changes from normal to the broken state. Thus, my findings will allow readers to detect anomalies in drill machines better and develop a more cost-effective system that can be conducted well on a small dataset. There is always background noise and acoustic noise in sounds, which affect the accuracy of the classification system. My hypothesis was that noise suppression methods would improve the sound classification application's accuracy. The result of my research is a sound separation method using short-time Fourier transform (STFT) frames with overlapped content. Unlike traditional STFT conversion, in which every sound is converted into one image, a different approach is taken. In contrast, splitting the signal into many STFT frames can improve the accuracy of model prediction by increasing the variability of the data. Images of these frames separated into clean and noisy ones are saved as images, and subsequently fed into a pre-trained CNN for classification. This enables the classifier to become robust to noise. The FSDNoisy18k dataset is chosen in order to demonstrate the efficiency of the proposed method. In experiments using the proposed approach, 94.14 percent of 21 classes were classified successfully, including 20 classes of sound events and a noisy class. / <p>Vid tidpunkten för disputationen var följande delarbeten opublicerade: delarbete 2 och 3 inskickat.</p><p>At the time of the doctoral defence the following papers were unpublished: paper 2 and 3 submitted.</p> / AISound – Akustisk sensoruppsättning för AI-övervakningssystem / MiLo — miljön i kontrolloopen