State estimation is essential to access observable Network models for online monitoring and analyzing of power systems. Due to the integration of distributed energy resources and new technologies, state estimation in distribution systems would be necessary. However, accurate input data are essential for an accurate estimation along with knowledge on the possible correlation between the real and pseudo measurements data. This study presents a new approach to model errors for the distribution system state estimation purpose. In this paper, pseudo measurements are generated using a couple of real measurements data by means of the artificial Neural Network method. In the proposed method, the radial basis function Network with the Gaussian kernel is also implemented to decompose pseudo measurements into several components. The robustness of the proposed error modeling method is assessed on IEEE 123-bus distribution test system where the problem is optimized by the imperialist competitive algorithm. The results evidence that the proposed method causes to increase in detachment accuracy of error components which results in presenting higher quality output in the distribution state estimation.

An adaptive version of growing and pruning RBF Neural Network has been used to predict the system output and implement Linear Model-Based Predictive Controller (LMPC) and Non-linear Model-based Predictive Controller (NMPC) strategies. A radial-basis Neural Network with growing and pruning capabilities is introduced to carry out on-line model identification. An Unscented Kalman Filter (UKF) algorithm with an exponential time-varying forgetting factor has been presented to enable the Neural Network model to track any time-varying process dynamic changes. An adaptive NMPC has been designed based on the sequential quadratic programming technique. The paper makes use of a dynamic linearization approach to extract a linear model at each sampling time instant so as to develop an adaptive LMPC. The servo and regulating performances of the proposed adaptive control schemes have been illustrated on a non-linear Continuous Stirred Tank Reactor (CSTR) as a benchmark problem. The simulation results demonstrate the capability of the proposed identification strategy to effectively identify compact, accurate and transparent model for the CSTR process. It is shown that the proposed adaptive NMPC controller presents better improvement with faster response time for both servo and regulatory control objectives in comparison with the proposed adaptive LMPC, an adaptive generalized predictive controller based on Recursive Least Squares (RLS) algorithm and well-tuned PID controllers.

Data acquisition of chemical engineering processes is expensive and the collected data are always contaminated with inevitable measurement errors. Efficient algorithms are required to filter out the noise and capture the true underlying trend hidden in the training data sets.egularization Networks, which are the exact solution of multivariate linear regularization problem, provide appropriate facility to perform such a demanding task. These Networks can be represented as a single hidden layer Neural Network with one neuron for each distinct exemplar. Efficient training of Regularization Network requires calculation of linear synaptic weights, selection of isotropic spread (s) and computation of optimum level of regularization (l*). The latter parameters (s and l*) are highly correlated with each other. A novel method is presented in this article for development of a convenient procedure for de-correlating the above parameters and selecting the optimal values of l* and. s* The plot of l* versus s suggests a threshold s* that can be regarded as the optimal isotropic spread for which the Regularization Network provides appropriate model for the training data set. It is also shown that the effective degrees of freedom of a Regularization Network are a function of both regularization level and isotropic spread. A readily calculable measure of the approximate degrees of freedom of a Regularization Network is also introduced which may be used to de-couple l* and s.

The Hartley transform, a real-valued alternative to the complex Fourier transform, is presented as an efficient tool for the analysis and simulation of earthquake accelerograms. This paper is introduced a novel method based on discrete Hartley transform (DHT) and radial basis function (RBF) Neural Network for generation of artificial earthquake accelerograms from specific target spectrums. Acceleration time histories of horizontal earthquake ground motion are obtained by the capability of learning of RBF Neural Network to expand the knowledge of the inverse mapping from the response spectrum to earthquake accelerogram. In the first step, Hartley transform is used to decompose earthquake accelerograms, then a RBF Neural Network is trained to learn to relate the response spectrum to Hartley spectrum. Finally, the generated accelerogram using inverse discrete Hartley transform is obtained from target spectrum. Approximately 200 uniformly scaled horizontal ground motion records from recent Iran’s earthquakes are used to decompose with real Hartley transform and train Networks.

The Hartley transform, a real-valued alternative to the complex Fourier transform, is presented as an efficient tool for the analysis and simulation of earthquake accelerograms.This paper is introduced a novel method based on discrete Hartley transform (DHT) and radial basis function (RBF) Neural Network for generation of artificial earthquake accelerograms from specific target spectrums. Acceleration time histories of horizontal earthquake ground motion are obtained by the capability of learning of RBF Neural Network to expand the knowledge of the inverse mapping from the response spectrum to earthquake accelerogram. In the first step, Hartley transform is used to decompose earthquake accelerograms, then a RBF Neural Network is trained to learn to relate the response spectrum to Hartley spectrum. Finally, the generated accelerogram using inverse discrete Hartley transform is obtained from target spectrum. Approximately 200 uniformly scaled horizontal ground motion records from recent Iran’s earthquakes are used to decompose with real Hartley transform and train Networks.