Introduction. Switched reluctance motor (SRM) is a type of electric motor featuring nonlinear magnetic characteristics. The flux linkage or inductance profile of SRM is usually required for the purpose of high control performance, and can be normally obtained through conventional static test by using DC or AC method when the rotor is locked. Problem. However, it is not practical to use the conventional method of measurement when the specific apparatus for locking the rotor is unavailable. Besides, due to the magnetic nonlinearity of SRM, the saturation effect makes it difficult to obtain the saturated magnetic characteristics, and the conventional static AC test fails to address this problem. Novelty. In this paper, a dynamic measurement method of the magnetization curves of SRM is proposed which allows the measurement take place while the motor is running with load. Methodology. Based on the conventional static AC test, the proposed measurement handles the saturation problem successfully by introducing a DC offset in the high frequency AC voltage. Phase inductance with different rotor positions and currents can be obtained by analyzing simple equivalent circuit. Practical value. Simulation is conducted in MATLAB/Simulink environment and the results have verified that the proposed dynamic measurement can effectively obtain the magnetic characteristics of SRM.

Introduction. The operation of switched reluctance motor requires prior knowledge of the rotor position, obtaining from either low resolution photocoupler based position sensor or high resolution shaft encoder, to control the on/off states of the power switches. Problem. However, using physical position sensor in harsh environment will inevitably reduce the reliability of the motor drive, in which sensorless control comes into play. Novelty. In this paper, a sensorless control scheme of switched reluctance motor is proposed. Methodology. The method is based on a simple analytical model of the flux-linkage curves rather than the conventional approach that normally uses a look-up table to store all the data points of the flux-linkage curves. By measuring the phase current, rotor position can be deduced from the analytical model. Practical value. Simulation results are given and the proposed sensorless scheme is verified to provide a moderate position estimation accuracy in a wide speed range in both unsaturated and saturated conditions.

Designing molecules that have desired properties is one of the challenging tasks of drug design. Among the many molecular generative models, a generative adversarial network (GAN), is able to generate molecule structures with desirable chemical properties via reinforcement learning. Generating valid molecules is the foremost task of any molecular generative model, since invalid molecules cannot be synthesized. We base our research on a molecular generative adversarial network (MolGAN) architecture to investigate how the validity score is influenced in different scenarios. First, we verify that the Vanilla GAN structure can produce valid molecules in measure, and that the reward network, along with Vanilla GAN, can further increase the validity score in a reinforcement learning manner. Then, the procedure for solely optimizing the validity score is tested, followed by an assessment of validity score maintenance while other chemical properties are being optimized. We found that multiple aspects, including loss functions, hyper parameters, and training sequences, must be carefully considered and optimized to raise the validity score of molecular generation alone or in concurrence with the optimizing of other chemical property scores.INDEX TERMS Drug design, molecular generation, generative adversarial network (GAN), molecular generative adversarial network (MolGAN).