Daily Temperature Profile Prediction for the District Heating Application

Application of artificial neural networks to assessment of ship manoeuvrability qualities This paper presents an attempt to applying neural networks for assessment of parameters of standard manoeuvrability tests, i.e. circulation test and zig-zag test. Methodological approach to application of neural networks as well as applied network structures and neuron activation functions are generally presented. Also, results of simulations performed by means of the elaborated networks are given in comparison with test cases selected at random. In order to analyze and reveal general trends, correlation relationships between results from network simulations and test cases were calculated and are presented as well.

This paper presents the practical application of cm artificial neural network (ANN) to the problem of concrete mix in a factory. After a brief introduction to the complex problem of concrete mixes design and a quick review of the fundamental basis of neurocomputation, an optimal neural network model has been developed to cope with such a problem. For training the net, several control mixes have been fabricated recording in all cases both the characteristic 28 days compressive strength and the workability measured in terms of the slump of the Abrams' cone. After the training process of the net, the power of its predictive ability is checked by comparison of the results obtained with those corresponding to four reference mixes; in this way, it is shown that the considered approach can be used in multicriterial search for optimal concrete mixes.

The application of artificial neural networks to the estimation of bioprocess variables will be discussed. In fermentation processes, direct on-line measurements of primary process variables usually are unavailable. The state of the cultivation, therefore, has to be inferred from measurements of secondary variables and any previous knowledge of process dynamics. This research investigates the learning, recall and generalization characteristics of neural networks trained to model the nonlinear behavior of a fermentation process. Two different neural network methodologies are discussed, namely, feed-forward and recurrent neural networks, which differ in their treatment of time dependence. The neural networks are trained by backpropagation using a conjugate gradient technique, which provides a dramatic improvement in the convergence speed. The objective is to use environmental and physiological information available from on-line sensors to estimate concentrations of species in a bioreactor. Results of the neural network estimators are presented, based on experimental data available from the ethanol production by Zymomonas mobilis fermentation. The feed-forward and recurrent neural network methodologies are demonstrated to perform suitably as unmeasurable state estimators. Both networks offer comparable abilities of recall, but recurrent networks perform better than feed-forward networks in generalization.

This paper introduces an application of artificial neural network use in Chinese coal mines and states its efficiency in mining problem solving. This paper provides a proposal for further development and application of the neural network in the coal mines and points out the technical problems which should receive attention.

This paper proposes the application of Artificial Neural Networks (ANN) in forecasting miners' occupational disease incidence, examines the effect of BP network topology and parameters on training time, describes a neural network that uses Legendre polynomials as nonlinear weights and tries to find the most suitable network model to solve the problem via analyzing the forecast result. Through computer simulation, it clarifies that this model could forecast miners' pneumoconiosis incidence commendably.

An experimental application of Artificial Neural Networks to Eating Disorders is presented. The sample, composed of 172 cases (all women) collected at the Centre for the Diagnosis and Treatment of Eating Disorders of the 1st Medical Division of the St. Eugenio Hospital of Rome, was subdivided, on the basis of the diagnosis made by the specialist of the St. Eugenio, into four classes: Anorexia Nervosa (AN), Nervous Bulimia (NB), Binge Eating Disorders (BED) and Psychogenic Eating Disorders that are Not Otherwise Specified (PED-NOS). The data base was composed of 124 different variables: generic information, alimentary behavior, eventual treatment and hospitalization, substance use, menstrual cycles, weight and height, hematochemical and instrumental examinations, psychodiagnostic tests, etc. The goal of this experiment was to verify the accuracy of the Neural Networks in recognising anorexic and bulimic patients. This article describes 6 experiments, using a Feed Forward Neural Network, each one using different variables. Starting from only the generic variables (life styles, family environment, etc.) and hematoclinical and instrumental examinations, a Neural Networks provided 86.94% of the prediction precision. This work is meant to be a first contribution to creating diagnostic procedures for Eating Disorders, that would be simple and easy-to-use by professionals who are neither psychologists nor psychiatrists nor psychotherapists but who are, however, among the first to meet these patients and who are therefore called upon to give such patients the very first pieces of advice on seeking proper treatment.

Development and application of reservoir models and artificial neural networks for optimizing ventilation air requirements in development mining of coal seams

The aim of the paper is to discuss the application of classification functions and artificial neural networks (such as multilayer perceptron and radial basis function) to recognize the risk category of investigated companies. The research is based on data from 295 enterprises that applied for credit in two regional banks operating in Poland. Each firm is described by 13 diagnostic variables and potential borrowers are classified into four classes. The efficiency of classification is evaluated in terms of classification errors calculated from the actual classification made by the credit officers. The results of the experiments show that application of artificial neural networks and classification functions can support the creditworthiness evaluation of borrowers.

The paper presents an innovative method for smoothing fluctuations of heat flux, using the thermal energy storage unit (TES Unit) with phase change material and Artificial Neural Networks (ANN) control. The research was carried out on a pilot large-scale installation, of which the main component was the TES Unit with a heat capacity of 500 MJ. The main challenge was to smooth the heat flux fluctuations, resulting from variable heat source operation. For this purpose, a molten salt phase change material was used, for which melting occurs at nearly constant temperature. To enhance the smoothing effect, a classical control system based on PID controllers was supported by ANN. The TES Unit was supplied with steam at a constant temperature and variable mass flow rate, while a discharging side was cooled with water at constant mass flow rate. It was indicated that the operation of the TES Unit in the phase change temperature range allows to smooth the heat flux fluctuations by 56%. The tests have also shown that the application of artificial neural networks increases the smoothing effect by 84%.

The Application of Artificial Neural Networks to the Identification of New Spinosoids with Improved Biological Activity toward Larvae of Heliothis virescens

The impact of process variations, operating conditions and environmental changes on high-speed system performances is becoming more significant due to the shrinking bit time and supply voltage. Consequently, accurate and exhaustive characterization is essential to improve the robustness and yield of current systems. However, the statistics of the system performance obtained from limited measurement can be unrepresentative of the final product. This paper describes the application of an artificial neural network (ANN) to generate the performance distribution and statistics of chip-to-chip communication system from measurements. A set of experiments is performed on a hardware system to capture the relationships between system parameters and performance. The measured data is then used to train the neural network that is used to efficiently generate realistic distributions and statistics of the system performance. The accuracy of the proposed approach is verified using a prototype system. The sensitivity of setup and hold times of the system is studied as functions of supply voltages and temperature. The results of the proposed method are also compared to those from measurements and regression models.

The liquid fuel quality control is an important issue that brings benefits for the State, for the consumers and for the environment. The conformity analysis, in special for gasoline, demands a rigorous sampling technique among gas stations and other economic agencies, followed by a series of standard physicochemical tests. Such procedures are commonly expensive and time demanding and, moreover, a specialist is often required to carry out the tasks. Such drawbacks make the development of alternative analysis tools an important research field. The fuel refractive index is an additional parameter to help the fuel conformity analysis, besides the prospective optical fiber sensors, which operate like transducers with singular properties. When this parameter is correlated with the sample density, it becomes possible to determine conformity zones that cannot be analytically defined. This work presents an application of artificial neural networks based on Radial Basis Function to determine these zones. A set of 45 gasoline samples, collected in several gas stations and previously analyzed according to the rules of Agência Nacional do Petróleo, Gás Natural e Biocombustíveis, a Brazilian regulatory agency, constituted the database to build two neural networks. The input variables of first network are the samples refractive indices, measured with an Abbe refractometer, and the density of the samples measured with a digital densimeter. For the second network the input variables included, besides the samples densities, the wavelength response of a long‐period grating to the samples refractive indices. The used grating was written in an optical fiber using the point‐to‐point technique by submitting the fiber to consecutive electrical arcs from a splice machine. The output variables of both Radial Basis Function Networks are represented by the conformity status of each sample, according to report of tests carried out following the American Society for Testing and Materials and/or Brazilian Association of Technical Rules standards. A subset of 35 samples, randomly chosen from the database, was used to design and calibrate (train) both networks. The two networks topologies (numbers of Radial Basis Function neurons of the hidden layer and function radius) were built in order to minimize the root mean square error. The subset composed by the other 10 samples was used to validate the final networks architectures. The obtained results have demonstrated that both networks reach a good predictive capability.

The paper considers the application of Artificial Neural Networks for the demodulation of SEFDM signals for BPSK and QPSK modulation schemes. Two approaches to building network architecture (classification and regression) and 12 learning methods were compared by BER performance. The network output size corresponds to the modulation order M in case of classification. The network output is decimal and always has one neuron in case of regression. An example of OFDM-QPSK signals showed that classification gives the best result. We also considered a different number of input neurons of the network. It is shown that the use of information from neighboring subcarriers of the SEFDM signal can significantly improve the BER performance. In this paper, neural network methods were opposed to the classical maximum likelihood algorithm and to the feedback algorithm. For the BPSK constellation, the largest gain was 6.8 dB and 2.3 dB, respectively (by level BER = 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">–3</sup> ). For the QPSK constellation, the probability of a bit error has decreased by a maximum of 7 times (by level <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$E_{b}/N_{0}=15\mathbf{dB})$</tex> .

This article presents an application of Artificial Neural Networks (ANNs) and multiple regression models for estimating mean annual maximum discharge (index flood) at ungauged sites. Both approaches were tested for 145 small basins in Slovakia in areas ranging from 20 to 300 km2. Using the objective clustering method, the catchments were divided into ten homogeneous pooling groups; for each pooling group, mutually independent predictors (catchment characteristics) were selected for both models. The neural network was applied as a simple multilayer perceptron with one hidden layer and with a back propagation learning algorithm. Hyperbolic tangents were used as an activation function in the hidden layer. Estimating index floods by the multiple regression models were based on deriving relationships between the index floods and catchment predictors. The efficiencies of both approaches were tested by the Nash-Sutcliffe and a correlation coefficients. The results showed the comparative applicability of both models with slightly better results for the index floods achieved using the ANNs methodology.

The paper present primary results on an application of artificial neural networks to predictions of the influence of technological parameters of continuous casting on the quality of casted ingots. Development of the classification method for ingots was the main objective of the work. The input data composed casting parameters and quality evaluation for 852 ingots. An analysis shows that development of models and classification methods, which will predict the quality of ingots, is possible. Beta versions of statistical and neuron models were developed. Iy is shown in the paper that an accuracy of the models depends on the quality assessment method. Neural network classification yielded better accuracy, about 96%. An attempt of development of one general classification method led to a decrease of predictions of the quality. Classification methods require further validation. The general conclusion from the work is that at this stage separate classification methods have to be created for each steel type and for each casting machine.