Case Study: Finite Element Method and Artificial Neural Network Models for Flow through Jeziorsko Earthfill Dam in Poland

A finite element method (FEM) and an artificial neural network (ANN) model were developed to simulate flow through Jeziorsko earthfill dam in Poland. The developed FEM is capable of simulating two-dimensional unsteady and nonuniform flow through a nonhomogenous and anisotropic saturated and unsaturated porous body of an earthfill dam. For Jeziorsko dam, the FEM model had 5,497 triangular elements and 3,010 nodes, with the FEM network being made denser in the dam body and in the neighborhood of the drainage ditches. The ANN model developed for Jeziorsko dam was a feedforward three layer network employing the sigmoid function as an activator and the back-propagation algorithm for the network learning. The water levels on the upstream and downstream sides of the dam were input variables and the water levels in the piezometers were the target outputs in the ANN model. The two models were calibrated and verified using the piezometer data collected on a section of the Jeziorsko dam. The water levels computed by the models satisfactorily compared with those measured by the piezometers. The model results also revealed that the ANN model performed as good as and in some cases better than the FEM model. This case study offers insight into the adequacy of ANN as well as its competitiveness against FEM for predicting seepage through an earthfill dam body.     

Numerical Simulation of Flood Wave Propagation in Two-Dimensions in Densely Populated Urban Areas due to Dam Break

Dams are important structures having many functions such as water supply, flood control, hydroelectric power and recreation. Although dam break failures are very rare events, dams can fail with little warning and the damage at the downstream of the dam due to the flood wave can be catastrophic. During a dam failure, immense volume of water is mobilized at very high speed in a very short time. The momentum of the flood wave can turn to a very destructive impact force in residential areas. Therefore, from risk point of view, understanding the consequences of a possible dam failure is critically important. This study deals with the methodology utilized for predicting the flood wave occurring after the dam break and analyses the propagation of the flood wave downstream of the dam. The methodology used in this study includes creation of bathymetric, DEM and land use maps; routing of the flood wave along the valley using a 1D model; and two dimensional numerical modeling of the propagation and spreading of flood wave for various dam breaching scenarios in two different urban areas. Such a methodology is a vital tool for decision-making process since it takes into account the spatial heterogeneity of the basin parameters to predict flood wave propagation downstream of the dam. Proposed methodology is applied to two dams; Porsuk Dam located in Eski?ehir and Alibey Dam located in Istanbul, Turkey. Both dams are selected based on the fact that they have dense residential areas downstream and such a failure would be disastrous in both cases. Model simulations based on three different dam breaching scenarios showed that maximum flow depth can reach to 5 m at the border of the residential areas both in Eski?ehir and in Istanbul with a maximum flow velocity of 5 m/s and flood waves having 0.3 m height reach to the boundary of the residential area within 1 to 2 h. Flooded area in Eski?ehir was estimated as 127 km², whereas in Istanbul this area was 8.4 km² in total.     

Predicting Mean and Bankfull Discharge from Channel Cross-Sectional Area by Expert and Regression Methods

This study employed four methods—non-linear regression, fuzzy logic (FL), artificial neural networks (ANNs), and genetic algorithm (GA)-based nonlinear equation—for predicting mean discharge and bank-full discharge from cross-sectional area. The data compiled from the literature were separated into two groups—training (calibration) and testing (verification). Using training data sets, the methods were calibrated to obtain optimal values of the coefficients of the non-linear regression method; optimal number of fuzzy subsets, their base widths and fuzzy rules for the fuzzy method; and the optimal number of neurons in the hidden layer, the learning rate and momentum factor values for the ANN model. The GA-based method employed 100 chromosomes in the initial gene pool, 80% cross over rate and 4% mutation rate in determining the optimal values of the coefficients of the constructed nonlinear equation. The calibrated methods were then applied to the test data sets. The test results showed that the non-linear regression, ANN and GA-based methods were comparable in predicting the mean discharge while the fuzzy method produced high errors and low accuracy. The GA-based method had the highest accuracy of 75%. In terms of predicting bankfull discharge, all methods produced satisfactory results, although the fuzzy method had the lowest accuracy of 33%. The results of sensitivity analysis, which is limited to the GA-based and nonlinear regression methods, showed that the GA-based method calibrated with low bankfull discharge values can be successfully applied to predict high bankfull discharge values. This has important implications for predicting bankfull rates at ungauged sites. On the other hand, the sensitivity analysis results also showed that both the non-linear regression and GA-based methods have poor extrapolation capability for predicting mean discharge data.     

Prediction of rainfall runoff‐induced sediment load from bare land surfaces by generalized regression neural network and empirical model

Based on three rainfall run‐off‐induced sediment transport data for bare surface experimental plots, the generalized regression neural network (GRNN) and empirical models were developed to predict sediment load. Rainfall intensity, slope, rainfall duration, soil particle median diameter, clay content of the soil, rill density and soil particle mass density constituted the input variables of the models while sediment load was the target output. The GRNN model was trained and tested. The GRNN model was found successful in predicting sediment load. Sensitivity analysis by the GRNN model revealed that slope and rainfall duration were the most sensitive parameters. In addition to the GRNN model, two empirical models were proposed: (1) in the first empirical model, all the input variables were related to the sediment load, and (2) in the second empirical model, only rainfall intensity, slope and rainfall duration were related to the sediment load. The empirical models were calibrated and validated. At the calibration stage, the coefficients and the exponents of the empirical models were obtained using the genetic algorithm optimization method. The validated empirical models were also applied to two more experimental data sets: (1) one data set was from a field experiment, and (2) one set was from a laboratory experiment. The results indicated the success of the empirical models in predicting sediment load from bare land surfaces.     

Experimental and artificial neural network modeling study on soot formation in premixed hydrocarbon flames

The formation of soot in premixed flames of methane, ethane, propane, and butane was studied at three different equivalence ratios. Soot particle sizes, number densities, and volume fractions were determined using classical light scattering measurement techniques. The experimental data revealed that the soot properties were sensitive to the fuel type and combustion parameter equivalence ratio. Increase in equivalence ratio increased the amount of soot formed for each fuel. In addition, methane flames showed larger particle diameters at higher distances above the burner surface and propane, ethane, and butane flames came after the methane flames, respectively. Three-layer, feed-forward type artificial neural networks having seven input neurons, one output neuron, and five hidden neurons for soot particle diameter predictions and seven hidden neurons for volume fraction predictions were used to model the soot properties. The network could not be trained and tested with sufficient accuracy to predict the number density due to a large data range and greater uncertainty in determination of this parameter. The number of complete data set used in the model was 156. There was a good agreement between the experimental and predicted values, and neural networks performed better when predicting output parameters (i.e. soot particle diameters and volume fractions) within the limits of the training data.     

Forecasting Ambient Air SO2 Concentrations Using Artificial Neural Networks

An Artificial Neural Networks (ANNs) model is constructed to forecast SO₂ concentrations in Izmir air. The model uses meteorological variables (wind speed and temperature) and measured particulate matter concentrations as input variables. The correlation coefficient between observed and forecasted concentrations is 0.94 for the network that uses all three variables as input parameters. The root mean square error value of the model is 3.60 μg/mt³. Considering the limited number of available input variables, model performances show that ANNs are a promising method of modeling to forecast ambient air SO₂ concentrations in Izmir.     

Supervised Intelligent Committee Machine Method for Hydraulic Conductivity Estimation

Hydraulic conductivity is the essential parameter for groundwater modeling and management. Yet estimation of hydraulic conductivity in a heterogeneous aquifer is expensive and time consuming. In this study; artificial intelligence (AI) models of Sugeno Fuzzy Logic (SFL), Mamdani Fuzzy Logic (MFL), Multilayer Perceptron Neural Network associated with Levenberg–Marquardt (ANN), and Neuro-Fuzzy (NF) were applied to estimate hydraulic conductivity using hydrogeological and geoelectrical survey data obtained from Tasuj Plain Aquifer, Northwest of Iran. The results revealed that SFL and NF produced acceptable performance while ANN and MFL had poor prediciton. A supervised intelligent committee machine (SICM), which combines the results of individual AI models using a supervised artificial neural network, was developed for better prediction of the hydraulic conductivity in Tasuj plain. The performance of SICM was also compared to those of the simple averaging and weighted averaging intelligent committee machine (ICM) methods. The SICM model produced reliable estimates of hydraulic conductivity in heterogeneous aquifers.     

Analysis of a combined power and cooling cycle for low‐grade heat sources

This paper presents a parametric analysis of a combined power/cooling cycle, which combines the Rankine and absorption refrigeration cycles, uses ammonia–water mixture as the working fluid and produces power and refrigeration, while power is the primary goal. This cycle, also known as the Goswami Cycle, can be used as a bottoming cycle using waste heat from a conventional power cycle or as an independent cycle using low‐temperature sources such as geothermal and solar energy. Optimum operating conditions were found for a range of ammonia concentration in the basic solution, isentropic turbine efficiency and boiler pressure. It is shown that the cycle can be optimized for net work, cooling output, effective first law and exergy efficiencies. The effect of rectification cooling source (external and internal) on the cycle output was investigated, and it was found that an internal rectification cooling source always produces higher efficiencies. When ammonia vapor is superheated after the rectification process, cycle efficiencies increase but cooling output decreases. Copyright © 2010 John Wiley & Sons, Ltd.     

Relational Logging Design Pattern

Observability and traceability of developed software are crucial to its success in software engineering. Observability is the ability to comprehend a system’s internal state from the outside. Monitoring is used to determine what causes system problems and why. Logs are among the most critical technology to guarantee observability and traceability. Logs are frequently used to investigate software events. In current log technologies, software events are processed independently of each other. Consequently, current logging technologies do not reveal relationships. However, system events do not occur independently of one another. With this perspective, our research has produced a new log design pattern that displays the relationships between events. In the design we have developed, the hash mechanism of blockchain technology enables the display of the logs’ relationships. The created design pattern was compared to blockchain technology, demonstrating its performance through scenarios. It has been determined that the recommended log design pattern outperforms blockchain technology in terms of time and space for software engineering observability and traceability. In this context, it is anticipated that the log design pattern we provide will strengthen the methods used to monitor software projects and ensure the traceability of relationships.