Machine learning and artificial neural networks have shown to be applicable in modeling and simulation of complex physical phenomena as well as creating surrogate models trained with physics-based simulation data for numerous applications that require good computational performance. In this article, we review widely the surrogate modeling concept and its applications in the electrical machine context. We present comprehensively a workflow for developing data-driven surrogate models including data generation with physics-based simulation and design of experiments, preprocessing of training data, and training and testing of the surrogates. We compare neural networks and gradient boosting decision trees in modeling and simulation of torque behavior of a permanent magnet synchronous machine together with selected design of experiments approaches with respect to surrogate accuracy and computational efficiency. In addition, an approach to utilizing domain knowledge to create a hybrid surrogate model in order to improve the surrogate accuracy is shown. The accuracy of the selected hybrid neural network was better than with the gradient boosting approach and was close to the finite element simulation, whereas its run-time efficiency was significantly better compared to the finite element simulation with a speed-up factor of over 2,000. In addition, combining the sampling methods provided better results than the selected methods alone.
A transient model of an induction machine (IM) is developed in this work using an artificial neural network (ANN) surrogate model. The model is suitable to be used for direct-on-line IMs. The finite-element (FE)-based model of IM is used to generate the training, validation, and testing datasets. Different inputs and model configurations are investigated to find an optimal solution in developing the transient model. The proposed transient model is suitable to be used in digital twin services since it can estimate the current and torque accurately in real time based on only voltage and measured shaft speed.
New methods to perform time series classification arise frequently and multiple state-of-the-art approaches achieve high performance on benchmark datasets with respect to accuracy and computation time. However, often the modeling procedures do not include proper validation but rather rely only on either external test dataset or one-level cross-validation. ATSC-NEX is an automated procedure that employs sequential model-based optimization together with nested cross-validation to build an accurate and properly validated time series classification model. It aims to find an optimal pipeline configuration that includes the selection of input type and settings, as well as model type and hyperparameters. The results of a case study in which a model for the identification of diesel engine type is developed, show that the algorithm can efficiently find a well-performing pipeline configuration. The comparison between ATSC-NEX and some state-of-the-art methods on several benchmark datasets shows that similar accuracy can be achieved.
Rotor bars are one of the most failure-critical components in induction machines. We present an approach for developing a rotor bar fault identification classifier for induction machines. The developed machine learning-based models are based on simulated electrical current and vibration velocity data and measured vibration acceleration data. We introduce an approach that combines sequential model-based optimization and the nested cross-validation procedure to provide a reliable estimation of the classifiers’ generalization performance. These methods have not been combined earlier in this context. Automation of selected parts of the modeling procedure is studied with the measured data. We compare the performance of logistic regression and CatBoost models using the fast Fourier-transformed signals or their extracted statistical features as the input data. We develop a technique to use domain knowledge to extract features from specific frequency ranges of the fast Fourier-transformed signals. While both approaches resulted in similar accuracy with simulated current and measured vibration acceleration data, the feature-based models were faster to develop and run. With measured vibration acceleration data, better accuracy was obtained with the raw fast Fourier-transformed signals. The results demonstrate that an accurate and fast broken rotor bar detection model can be developed with the presented approach.
A robust open-source cloud-based workflow is developed for finite element (FE) data generation for active learning (AL) -based surrogate modelling. Special attention is paid to making the FE solution procedure as robust and fast as possible without human intervention by, e.g., implementing special convergence criteria, reliable parallel computation, and variable timestep length. In AL, a surrogate model automatically improves itself by iteratively querying more FE data. Using AL and large datasets generated with parallelised cloud FE simulations, we develop a surrogate model to rapidly predict induction machine steady-state torque, torque ripple, total losses, and current harmonic distortion, as a function of motor frequency, voltage, and slip. Results show that AL performs better than grid sampling and on average works as well as random sampling, but with some outputs, the results vary less with AL. In addition, accurate ripple estimation requires a much larger training dataset than the other variables.