Nonlinear Complex Model Research Articles

Nonlinear controller for Doubly Fed Induction Motor with bi-directional AC/DC/AC converter

This paper deals with the problem of controlling doubly-fed induction machines (DFIM). A nonlinear model of the whole controlled system, including the DFIM and the associated AC/DC rectifier and DC/AC inverter, is developed within the Park coordinates. A multi-loop nonlinear controller is developed, using the backstepping design technique. The controller is formally shown to meet its objectives i.e. accurate motor speed-reference tracking, tight regulation of the DC link voltage and power factor correction. The theoretical developments and results are validated by simulation on a representative nonlinear complex model.

Nanoindentation of hyperelastic polymer layers at finite deformation and parameter re-identification

Thin polymer layers on substrates have a wide range of application in important areas. However, it is impossible to measure the mechanical properties with the traditional testing methods. Recently, nanoindentation became a new but primary testing technique of thin layers. In the present work, based on a finite element model of contact mechanics and hyperelastic materials, nanoindentation of polymer layers is simulated with the finite element code ABAQUS®. Three often used hyperelastic models, that is, the neo-Hookean, Mooney–Rivlin and Yeoh models are investigated. The behaviour of these three models is compared to each other in different boundary value problems of nanoindentation in order to get some feeling of the different behaviour of various hyperelastic models under nanoindentation. In contrast to the traditional analytical method, the penetration depth is not restrained to avoid the influence of the substrate. A parameter re-identification strategy is employed to extract the parameters of the material models at small and finite deformation based on the principle of biological evolution. Furthermore, it is investigated how large the penetration depth has to be chosen in order to distinguish different models in reference to the load–displacement curves. Finally, the possibility is discussed of describing the data obtained by a non-linear complex model using the relatively simple approach based on the neo-Hookean model.

Entropía aproximada del efecto placebo en ensayos clínicos con antidepresivos de nueva generación

Recent research concluded that antidepressant drugs are ineffective in treating moderate or severe depression. Statistically, there are no differences between the results with active drugs or placebo. Some authors have attributed this fail to variability or irregularity of the placebo effect in depressed patients or artifacts induced by meta-analysis.This fact highlights the difficulties faced by the research of psychoactive drugs in depression and revives the debate about the usefulness of the employ of placebo in these studies. This study aimed to determine the variability of the placebo effect in antidepressant clinical trials in simple linear and nonlinear complex models. We performed a secondary analysis of data from 35 trials presented as evidence to the Food and Drug Administration (FDA) for approval of new generation antidepressants, all double-blind controlled with placebo in patients with unipolar mild or moderate depressive disorder, according to the criteria of "Diagnostic and Statistical Manual of Mental Disorders (DSM).Articles reviewed included 5133 patients . We have calculated the coefficient of variability, autocorrelation and approximate entropy of the placebo and treatment effects to determine whether the variability or regularity between different studies should be attributed to meta-analytical methods, placebo effect or ineffective treatment itself. The coefficient of variability in the placebo group was 26.49% and 18.81% in the treatment group. The placebo effect autocorrelation was within the confidential limits while the treatment group was outside showed cyclical variation. The approximate entropy value (ApEn N=35,m=2,R=2) in the placebo group was 0.5579 and 0.5744 in treatment group, leading to the conclusion that placebo effect is highly consistent and regular in complex non-linear models. The apparent variability of the placebo effect in depressed patients slould be due to artifacts induced by simple linear models analysis.