Induction Motor Modeling Based on a Fuzzy Clustering Multi-Model—A Real-Time Validation

This paper discusses a comparative study of two modeling methods based on multimodel approach. The first is based on C-means clustering algorithm and the second is based on K-means clustering algorithm. The two methods are experimentally applied to an induction motor. The multimodel modeling consists in representing the IM through a finite number of local models. This number of models has to be initially fixed, for which a subtractive clustering is necessary. Then both C-means and K-means clustering are exploited to determine the clusters. These clusters will be then exploited on the basis of structural and parametric identification to determine the local models that are combined, finally, to form the multimodel. The experimental study is based on MATLAB/SIMULINK environment and a DSpace scheme with DS1104 controller board. Experimental results approve that the multimodel based on K-means clustering algorithm is the most efficient.

A new multi-model coding algorithm based on MBE and spectral amplitude correlation between successive frames

The researches on spectral amplitude correlation between successive frames based on MBE (Multi-Band Excitation) are presented and a new MBE coding algorithm of muitimodels is proposed. It has been shown that the magnitude spectrum in MBE coding algorithm can be estimated in very high accuracy by using a simple linear predictive model of the spectral amplitude correlation between successive frames. This model employs oniy one gain coefficient and one predictive coefficient. The accuracy of estimated magnitude spectrum in this model is higher than that of the conventional MBE algorithm using the 10-th order LPC (Linear Predictive Coding) Inodel without the spectral amplitude correlation between successlve frames.This model is superior to the conventional MBE coding model in volced speech, but it is iess effective in unvoiced speech, so that the conventional MBE coding algorithm is still used for unvoiced speech. The new multi-model coding algorithm improves the quality of coding speech,and obtains good results

Multi-Model Communication and Data Assimilation for Mitigating Model Error and Improving Forecasts

Models for weather and climate prediction are complex, and each model typically has at least a small number of phenomena that are poorly represented, such as perhaps the Madden-Julian Oscillation(MJO for short) or El Nino-Southern Oscillation(ENSO for short) or sea ice. Furthermore, it is often a very challenging task to modify and improve a complex model without creating new deficiencies. On the other hand, it is sometimes possible to design a low-dimensional model for a particular phenomenon, such as the MJO or ENSO, with significant skill, although the model may not represent the dynamics of the full weather-climate system. Here a strategy is proposed to mitigate these model errors by taking advantage of each model’s strengths. The strategy involves inter-model data assimilation, during a forecast simulation, whereby models can exchange information in order to obtain more faithful representations of the full weather-climate system. As an initial investigation, the method is examined here using a simplified scenario of linear models, involving a system of stochastic partial differential equations(SPDEs for short)as an imperfect tropical climate model and stochastic differential equations(SDEs for short) as a low-dimensional model for the MJO. It is shown that the MJO prediction skill of the imperfect climate model can be enhanced to equal the predictive skill of the lowdimensional model. Such an approach could provide a route to improving global model forecasts in a minimally invasive way, with modifications to the prediction system but without modifying the complex global physical model itself.

A new approach for Bayesian model averaging

Bayesian model averaging(BMA) is a recently proposed statistical method for calibrating forecast ensembles from numerical weather models.However,successful implementation of BMA requires accurate estimates of the weights and variances of the individual competing models in the ensemble.Two methods,namely the Expectation-Maximization(EM) and the Markov Chain Monte Carlo(MCMC) algorithms,are widely used for BMA model training.Both methods have their own respective strengths and weaknesses.In this paper,we first modify the BMA log-likelihood function with the aim of removing the addi-tional limitation that requires that the BMA weights add to one,and then use a limited memory quasi-Newtonian algorithm for solving the nonlinear optimization problem,thereby formulating a new approach for BMA(referred to as BMA-BFGS).Several groups of multi-model soil moisture simulation experiments from three land surface models show that the performance of BMA-BFGS is similar to the MCMC method in terms of simulation accuracy,and that both are superior to the EM algo-rithm.On the other hand,the computational cost of the BMA-BFGS algorithm is substantially less than for MCMC and is al-most equivalent to that for EM.