Neural network-based failure rate prediction for De Havilland Dash-8 tires

An artificial neural network (ANN) model for predicting the failure rate of De Havilland Dash-8 airplane tires utilizing the two-layered feed-forward back-propagation algorithm as a learning rule is developed. The inputs to the neural network are independent variables and the output is the failure rate of the tires. Six years of data are used for model building and validation. Model validation, which reflects the suitability of the model for future prediction is performed by comparing the predictions of the model with that of Weibull regression model. The results show that the failure rate predicted by the ANN is closer in agreement with the actual data than the failure rate predicted by the Weibull model.     

Aggregation of Induction Motor Loads for Transient Stability Studies

This study investigates the dynamic representation of a group of induction motors by one or more equivalent motors. The unknown parameters of the equivalent are estimated through a weighted-least-squares procedure. Verification of simulated results with experimental ones suggests that a fifth order dynamic model of the aggregate with a ninth order parameter model is appropriate. The least squares technique was found to provide good convergence characteristics subject to satisfaction of the following conditions good initial estimate of the parameters, proper selection of acceleration factors and homogenity of the motor group.     

Software project management with GAs

A Project Scheduling Problem consists in deciding who does what during the software project lifetime. This is a capital issue in the practice of software engineering, since the total budget and human resources involved must be managed optimally in order to end in a successful project. In short, companies are principally concerned with reducing the duration and cost of projects, and these two goals are in conflict with each other. In this work we tackle the problem by using genetic algorithms (GAs) to solve many different software project scenarios. Thanks to our newly developed instance generator we can perform structured studies on the influence the most important problem attributes have on the solutions. Our conclusions show that GAs are quite flexible and accurate for this application, and an important tool for automatic project management.     

Modeling of multi gas-solid reactions; effect of bulk environment parameters on solid conversion

A unified gas-solid reaction model that considers multi gas-solid components is developed for reactions in porous pellets with multiple reactants. The transient variables and properties of the reactants and transport parameters are continually updated in the simulation. Temperature, pressure, solid conversions, and concentration profiles are well predicted. Structural changes are also considered. Heat, mass, and reaction front tracking equations for all the components are solved simultaneously through a combination of solution techniques. Illustrative validation results for the model are presented for hematite/nickel oxide reduction at 608 K. The importance of bulk temperature and pressure on the overall and individual grain conversions is explored for a range of interest. The response of system parameters to dynamic changes in boundary conditions is studied.     

FTIR,DTA and XRD study of sphene (CaTiSiO5) crystallization in a ceramic frit and a non-borate base glass

The primary objective of this study has been the application of Fourier transform infrared (FTi.r.) absorption spectroscopy for both qualitative and quantitative characterization of sphene (CaTiSiO₅) crystallization in test materials; namely, a CaO-TiO₂-B₂O₃ bearing ceramic frit-S and a similar non-borate base glass-S. Differential thermal analysis (DTA), X-ray diffraction (XRD) and scanning electron microscope/electron probe X-ray microanalysis (SEM/EPMA) techniques have also been used. FTi.r. absorption spectra have been shown to be capable of providing both qualitative and quantitative characterizations of crystal nucleation and growth in a frit-S and glass-S, being annealed between 800–1100° C. CaTiSiO₅ appears as the dominant phase and-cristobalite as the transitional phase in frit-S; whereas, -CaSiO₃ is dominant, CaTiSiO₅ being a minor phase in the non-borate glass-S. As given by DTA data, the intense stage of crystal growth for frit-S is about 120–125° C lower than that of glass-S. B₂O₃ content and the relative amounts of CaO and TiO₂ in the test specimens have been shown to give different modes of phase evolution and the onset temperature of nucleation. The activation energies,E _c, of crystal nucleation/growth was estimated by two different methods, namely, via DTA data and FTi.r. absorption spectra under the dominant surface nucleation mode for powder pellet specimens.E _c for CaTiSiO₅, -CaSiO₃ and-cristobalite in the frit-S and the non-borate base glass-S were estimated to be 219.6, 107.2 and 51.5 kJ mol^–1, respectively, parallel to the decreasing order of chemical complexity of the glass-forming system. Similar quantitative FTi.r. studies in the determination ofE _c for a broader scope of glass compositions, and compared with that based on XRD and DTA data, are to be encouraged so that the application of FTi.r. spectroscopy in glass-ceramics may be advanced.     

Study of delay-independent decentralized guaranteed cost control for large scale systems

This study introduces delay independent decentralized guaranteed cost control design method based on two controller structures for nonlinear uncertain interconnected large scale systems with time delays. First, a set of equivalent Takagi-Sugeno (T-S) fuzzy models are extended to represent the systems. Then a decentralized state-feedback guaranteed cost performance controller is proposed for the fuzzy systems. Based on delay independent Lyapunov functional approach, some sufficient conditions for the existence of the controller can be cast into the feasible problem of LMIs irrespective of the sizes of the time delays so that the system can be asymptotically stabilized for all considered uncertainties whose sizes are not larger than their bounds. Finally, the minimizing approach is proposed to search the suboptimal upper bound value of guaranteed cost function. Moreover, the corresponding conditions are extended into the generalized dynamic output-feedback close-loop system. Finally, the better control performances of the proposed methods are shown by the simulation examples.     

A rough set approach to feature selection based on power set tree

Feature selection is viewed as an important preprocessing step for pattern recognition, machine learning and data mining. Traditional hill-climbing search approaches to feature selection have difficulties to find optimal reducts. And the current stochastic search strategies, such as GA, ACO and PSO, provide a more robust solution but at the expense of increased computational effort. It is necessary to investigate fast and effective search algorithms. Rough set theory provides a mathematical tool to discover data dependencies and reduce the number of features contained in a dataset by purely structural methods. In this paper, we define a structure called power set tree (PS-tree), which is an order tree representing the power set, and each possible reduct is mapped to a node of the tree. Then, we present a rough set approach to feature selection based on PS-tree. Two kinds of pruning rules for PS-tree are given. And two novel feature selection algorithms based on PS-tree are also given. Experiment results demonstrate that our algorithms are effective and efficient.