Training and optimization of an artificial neural network controlling a hybrid power filter

A hybrid power compensator (HPC) consisting of a static VAr compensator and a dynamic compensator needs to be optimally controlled during the compensation of nonlinear loads. The HPC must be controlled to meet minimum requirements in terms of power factor and harmonic distortion, while at the same time minimizing its total cost. An artificial neural network (ANN) is used to control the HPC amidst a very dynamic power system environment. The performance of a reference ANN is evaluated while controlling an HPC connected to a typical nonlinear industrial load. The training and performance of the ANN is then optimized in terms of training set size, training set packing and ANN topology and the performance compared to the reference ANN. This paper highlights the importance of optimising the mentioned ANN parameters to achieve optimum ANN training and modeling accuracy. The results obtained reveals that the application of an ANN in controlling an HPC is feasible given that the ANN parameters are chosen appropriately.     

Start-up and steady-state results of a full-scale UASB reactor treating malting wastewater.

An Imhoff tank was reconstructed into a 250 m3 UASB reactor in order to treat a malting plant wastewater. The UASB was inoculated with sludge from an anaerobic lagoon used for slaughterhouse wastewater treatment. After two months of operation the reactor achieved full load with an HRT of 17 h, a COD removal higher than 80% and a biogas production of 300 m3/day (77% average methane content), with an organic loading rate of 3.6 kgCOD/m3.d (0.24 kgCOD/kgVSS.d). A yield coefficient of 0.09 gVSS/gCODrem was found from a mass balance. The fat present in the inoculated sludge (48 mg/gSSV) did not affect the start up performance. Sludge from the inoculum with high content of fat (270 mg/gSSV), was separated by flotation in the first week of operation. The COD removal efficiency was scarcely influenced by the reactor operation temperature (17-25 degrees C).     

Towards intelligent dressing

The aim of this article is to state the principles of an intelligent monitoring and control system for the grinding machine, comprising the dressing process as well as grinding stability.     

Radiative lifetimes of electrons and holes in a thin semiconductor layer

Radiative lifetimes of nondegenerate electrons and holes distributed uniformly in a semiconductor layer either deposited on a substrate or bounded by two dielectric media are calculated. The obtained expression takes into account the radiation reabsorbtion and interference effects and determines the dependences of the radiative lifetimes on the refractive indices of the external media and on the layer thickness.     

LiNbO3 optical waveguides formed in a new proton source

Proton-exchanged planar waveguides have been fabricated on Z-cut and X-cut lithium niobate crystals by using a new proton source formed by a mixture of benzoic and adipic acids. Waveguide index profiles and optical characteristics have been obtained at different values of the adipic-benzoic acid concentration ratio. The samples have been structurally characterized by Raman and infrared (IR) absorption spectroscopy and double-crystal X-ray diffraction. Good quality samples have been fabricated by using 30 mol% ratio dilution, showing very low scattering levels (<0.1 dB/cm), relatively high electrooptic coefficient (r₃₃=0.88 pm/V), and low relative percentage of interstitial protons (26%). The main factor limiting the waveguide optical properties is the substitutional-interstitial proton ratio, which can be easily controlled to produce good quality waveguides. A demonstration of the repeatability of the exchange process in the acid mixture is also provided     

Application of neural networks and machine learning in networkdesign

Communication network design is becoming increasingly complex, involving making networks more usable, affordable, and reliable. To help with this, we have proposed an expert network designer (END) for configuring, modeling, simulating, and evaluating large structured computer networks, employing artificial intelligence, knowledge representation, and network simulation tools. We present a neural network/knowledge acquisition machine-learning approach to improve the END's efficiency in solving the network design problem and to extend its scope to acquire new networking technologies, learn new network design techniques, and update the specifications of existing technologies