Modeling of a sequencing batch reactor treating municipal wastewater using multi-layer perceptron and radial basis function artificial neural networks

A sequencing batch reactor was modeled using multi-layer perceptron and radial basis function artificial neural networks (MLPANN and RBFANN). Then, the effects of influent concentration (IC), filling time (FT), reaction time (RT), aeration intensity (AI), SRT and MLVSS concentration were examined on the effluent concentrations of TSS, TP, COD and NH₄⁺-N. The results showed that the optimal removal efficiencies would be obtained at FT of 1 h, RT of 6 h, aeration intensity of 0.88 m³/min and SRT of 30 days. In addition, COD and TSS removal efficiencies decreased and TP and NH₄⁺-N removal efficiencies did not change significantly with increases of influent concentration. The TSS, TP, COD and NH₄⁺-N removal efficiencies were 86%, 79%, 94% and 93%, respectively. The training procedures of all contaminants were highly collaborated for both RBFANN and MLPANN models. The results of training and testing data sets showed an almost perfect match between the experimental and the simulated effluent of TSS, TP, COD and NH₄⁺-N. The results indicated that with low experimental values of input data to train ANNs the MLPANN models compared to RBFANN models are more precise due to their higher coefficient of determination (R²) and lower root mean squared errors (RMSE) values.     

The photocatalytic degradation of carbamazepine and prediction by artificial neural networks

The three layer artificial neural network model was applied to predict the degradation efficiency for carbamazepine in photocatalytic oxidation under UV radiation. Titania–zirconia was employed as a catalyst for the photooxidation. The catalyst was prepared using titanium isopropoxide and zirconium oxychloride by sol–gel method and characterized by transmission electron microscopy and BET analysis. Different process parameters such as, initial concentration of carbamazepine, pH of the solution, catalyst concentration and time of UV irradiation were employed as the input to the artificial neural network model and the output of the network was degradation efficiency of carbamazepine. The multilayer feed-forward networks with the Levenberg–Marquardt (trainlm) backpropagation training algorithm was used for the network training. The smallest mean square error was obtained for three-layer network with ‘logsig’ transfer function and five neurons in the hidden layer gave optimal results. A comparison between the predicted values and selective experimental data of degradation efficiency showed a high correlation coefficient (R²) of 0.997.     

A beetle antennae search improved BP neural network model for predicting multi-factor-based gas explosion pressures

A gas explosion in an underground structure may cause serious damage to the human body and ground buildings and may result in huge economic losses. The pressure of the gas explosion is an important parameter in determining its severity and designating an emergency plan. However, existing empirical and computational fluid dynamics (CFD) methods for pressure prediction are either inaccurate or inefficient when considering multiple influencing factors and their interrelationships. Therefore, for a more efficient and reliable prediction, the present study developed a multifactorial prediction model based on a beetle antennae search (BAS) algorithm improved back propagation (BP) neural network. A total of 317 sets of data which considered factors of geometry, gas, obstacle, vent, and ignition were collected from previous studies. The results showed that the established model can predict pressures accurately by low RMSE (43.4542 and 50.7176) and MAPE (3.9666% and 4.9605%) values and high R² (0.7696 and 0.7388) values for training and testing datasets, respectively. Meanwhile, the BAS algorithm was applied to improve both the calculation efficiency and the accuracy of the proposed model by enabling a more intelligent hyperparameter tuning method. Furthermore, the permutation importance of input variables was investigated, and the length (L) and the ratio of length and diameter (L/D) of geometry were found to be the most critical factors that affect the explosion pressure level.     

Prediction of gas explosion pressures: A machine learning algorithm based on KPCA and an optimized LSSVM

A gas explosion, as a common accident in public life and industry, poses a great threat to the safety of life and property. The determination and prediction of gas explosion pressures are greatly important for safety issues and emergency rescue after an accident occurs. Compared with traditional empirical and numerical models, machine learning models are definitely a superior approach. However, the application of machine learning in gas explosion pressure prediction has not reached its full potential. In this study, a hybrid gas explosion pressure prediction model based on kernel principal component analysis (KPCA), a least square support vector machine (LSSVM), and a gray wolf optimization (GWO) algorithm is proposed. A dataset consisting of 12 influencing factors of gas explosion pressures and 317 groups of data is constructed for developing and evaluating the KPCA-GWO-LSSVM model. The results show that the correlations among the 12 influencing factors are eliminated and dimensioned down by the KPCA method, and 5 composite indicators are obtained. The proposed KPCA-GWO-LSSVM hybrid model performs well in predicting gas explosion pressures, with coefficient of determination (R²), root mean square error (RMSE), and mean absolute error (MAE) values of 0.928, 26.234, and 12.494, respectively, for the training set; and 0.826, 25.951, and 13.964, respectively, for the test set. The proposed model outperforms the LSSVM, GWO-LSSVM, KPCA-LSSVM, beetle antennae search improved BP neural network (BAS-BPNN) models and reported empirical models. In addition, the sensitivity of influencing factors to the model is evaluated based on the constructed database, and the geometric parameters X1 and X2 of the confined structure are the most critical variables for gas explosion pressure prediction. The findings of this study can help expand the application of machine learning in gas explosion prediction and can truly benefit the treatment of gas explosion accidents.     

Traffic accident reconstruction and an approach for prediction of fault rates using artificial neural networks: A case study in Turkey

Objective: Currently, in Turkey, fault rates in traffic accidents are determined according to the initiative of accident experts (no speed analyses of vehicles just considering accident type) and there are no specific quantitative instructions on fault rates related to procession of accidents which just represents the type of collision (side impact, head to head, rear end, etc.) in No. 2918 Turkish Highway Traffic Act (THTA 1983). The aim of this study is to introduce a scientific and systematic approach for determination of fault rates in most frequent property damage–only (PDO) traffic accidents in Turkey.

Methods: In this study, data (police reports, skid marks, deformation, crush depth, etc.) collected from the most frequent and controversial accident types (4 sample vehicle–vehicle scenarios) that consist of PDO were inserted into a reconstruction software called vCrash. Sample real-world scenarios were simulated on the software to generate different vehicle deformations that also correspond to energy-equivalent speed data just before the crash. These values were used to train a multilayer feedforward artificial neural network (MFANN), function fitting neural network (FITNET, a specialized version of MFANN), and generalized regression neural network (GRNN) models within 10-fold cross-validation to predict fault rates without using software. The performance of the artificial neural network (ANN) prediction models was evaluated using mean square error (MSE) and multiple correlation coefficient (R).

Results: It was shown that the MFANN model performed better for predicting fault rates (i.e., lower MSE and higher R) than FITNET and GRNN models for accident scenarios 1, 2, and 3, whereas FITNET performed the best for scenario 4. The FITNET model showed the second best results for prediction for the first 3 scenarios. Because there is no training phase in GRNN, the GRNN model produced results much faster than MFANN and FITNET models. However, the GRNN model had the worst prediction results. The R values for prediction of fault rates were close to 1 for all folds and scenarios.

Conclusions: This study focuses on exhibiting new aspects and scientific approaches for determining fault rates of involvement in most frequent PDO accidents occurring in Turkey by discussing some deficiencies in THTA and without regard to initiative and/or experience of experts. This study yields judicious decisions to be made especially on forensic investigations and events involving insurance companies. Referring to this approach, injury/fatal and/or pedestrian-related accidents may be analyzed as future work by developing new scientific models.     

Artificial neural network modelling of As(III) removal from water by novel hybrid material

The present study reported a method for removal of As(III) from water solution by a novel hybrid material (Ce-HAHCl). The hybrid material was synthesized by sol–gel method and was characterized by XRD, FTIR, SEM–EDS and TGA–DTA. Batch adsorption experiments were conducted as a function of different variables like adsorbent dose, pH, contact time, agitation speed, initial concentration and temperature. The experimental studies revealed that maximum removal percentage is 98.85 at optimum condition: pH = 5.0, agitation speed = 180 rpm, temperature = 60 °C and contact time = 80 min using 9 g L⁻¹ of adsorbent dose for initial As(III) concentration of 10 mg L⁻¹. Using adsorbent dose of 10 g L⁻¹, the maximum removal percentage remains same with initial As(III) concentration of 25 mg L⁻¹ (or 50 mg L⁻¹). The maximum adsorption capacity of the material is found to be 182.6 mg g⁻¹. Subsequently, the experimental results are used for developing a valid model based on back propagation (BP) learning algorithm with artificial neural networking (BP-ANN) for prediction of removal efficiency. The adequacy of the model (BP-ANN) is checked by value of the absolute relative percentage error (0.293) and correlation coefficient (R² = 0.975). Comparison of experimental and predictive model results show that the model can predict the adsorption efficiency with acceptable accuracy.     

Coupling of polyhydroxyalkanoate production with volatile fatty acid from food wastes and excess sludge

In this paper, the synthesis of polyhydroxyalkanoates (PHAs) by activated sludge with aerobic dynamic feeding process was conducted in a sequencing batch reactor by using food wastes and excess sludge fermentation liquid as the carbon source. The volatile fatty acids (VFAs) in the fermentation liquid were divided into even-numbered (acetic and butyric acid) and odd-numbered (propionic and valeric acid). The experiments conducted by central-composite design (CCD) showed that the pH could significantly affect the ratio of even-numbered to odd-numbered VFAs. Statistical analysis indicated a positive correlation (R² = 0.97, P < 0.05) between the consumption of even-numbered VFAs and the synthesized of PHB, while the consumption of odd-numbered VFAs were correlated with the synthesized PHV. By controlling the ratio of even-numbered to odd-numbered VFAs, the contents of PHV in the PHAs could be controlled within the range of 22–30%. When fermentative VFAs were used as the substrate for the synthesis of PHAs, the microbial synthesis of PHA and biomass was higher than that mixture of analytically pure acids was used. These results are of vital significance for the comprehensive utilization of solid wastes.     

Anthropometric evaluation of cockpit designs

The objective of this research is to evaluate all the critical reaches in a cockpit and determine the visual sufficiency of a cockpit to accommodate 90% of potential pilots. While mismatches of measurements with cockpit dimensions are revealed, proposals are made to improve cockpit ergonomics. Regression models were generated to predict and assure adequate exterior vision. Mean, lower and upper control limits of all measurements were found acceptable except eye level. There are very strong positive relationships between stature and eye level (R²?=?0.972, p?R²?=?0.994, p?SD smaller than the eye level mean or seating adjustment limits in height may be changed. In general, cockpit design is acceptable in terms of fit/reach accommodation for pilots, except eye level and visual variables that could be solved by better seat adjustments.     

Prediction of ultimate bearing capacity of Tubular T-joint under fire using artificial neural networks

An artificial neural network (ANN) model is developed for the prediction of the ultimate bearing capacity of tubular T-joint under fire. The input parameters of the network are composed of the diameter ratio (β), the wall thickness ratio (τ), the diameter–thickness ratio (γ) and the temperature (T). The output parameter is composed of the ultimate bearing capacity. In this paper, the training and testing data of the neural network are obtained using the finite element program ABAQUS. The network is trained by 216 dataset and tested by 27 dataset. In the process of training of the network, the Levenberg-Marquardt back-propagation algorithm is adopted. The ‘tansig’ function is adopted in the hidden layer, and the ‘purelin’ function is adopted in the output layer. The results predicted by ANN are compared with the results simulated by finite element method (FEM). These results show that the prediction of the ultimate bearing capacity using the network model is accurate and effective.     

基于PCA-GA-BP神经网络的瓦斯含量预测分析

为提高煤层瓦斯含量预测的效率和准确率,提出了先采用主成份分析（PCA）方法来降低变量间的相关性,然后将遗传算法（GA）与BP神经网络相结合的煤层瓦斯含量预测的新方法。为了避免BP神经网络收敛速度慢、易陷入局部极小值等问题,算法采用GA对BP神经网络的权值和阈值进行优化,利用Matlab软件进行编程,建立了BP神经网络和GA-BP神经网络瓦斯含量预测模型。选取淮南某矿瓦斯含量及其影响因素作为实验数据对该模型进行了实例分析,将主成份回归和BP网络算法预测结果与该模型进行了对比分析。结果表明:PCA-GA-BP网络预测模型平均相对误差为2.759%,预测效果明显优于主成份回归和BP网络预测模型,可以准确的预测煤层瓦斯含量。     

BLEVE: Safety distances estimation by simple models based on the Jakob number

One of the most important points in the design of inherently safe processes is to estimate reliable distances among process units at preliminary stages of the plant project to minimize losses and damages caused by the potential occurrence of technological accidents. Therefore, in this paper the achievement of simple, general, dimensionless and reliable equations (Simple Dimensionless Models SDMs) for the direct estimation of safety distances considering the occurrence of BLEVE (Boiling Liquid Expanding Vapour Explosion) event, is proposed. The developed models directly relate safety distances with critical design/operation variables (involved substance, vessel volume, target vulnerability and explosion temperature), which are easily accessible at early stages of the plant project. SDMs are achieved by analysing the influence of these simple variables on the safety distances, which are estimated using a selected rigorous model (Reference Model R_fM). This task is simplified by the incorporation of the Jakob Number as an input variable, allowing to obtain dimensionless models and simultaneously an adequate representation of the explosion conditions and the involved substances. As result, the achieved SDMs demonstrate a particularly good fit with respect to the R_fM estimations and, at the same time, reliability and versatility. As it is shown in the analysed study cases (involving critical decision variables for the process design and the system safety), the SDMs prove to be also accurate, general, and easily incorporable into more complex optimization problems (QRA analysis, design of emergency plans, safety distance estimation to minimize the probability of domino effects, optimal layout designs, among others).     

集合卡曼滤波算法对FARSITE模型林火蔓延预测的修正效果研究

针对火源位置输入偏差导致的FARSITE林火行为模型火线预测不准确的问题,提出了一种基于集合卡曼滤波算法的动态修正方法。利用FARSITE对复杂工况下的林火蔓延过程进行数值模拟,以火线位置为待修正参量,以均方根误差(RMSE)为评价指标,对算法的可行性进行了验证,并研究了算法的集合元素个数,观测数据标准差及同化频率对FARSITE预测偏差的修正效果的影响。结果表明:算法能显著提高FARSITE火线预测精度;逐时同化时:集合元素个数为5 时,算法的修正效果并不理想,随着集合元素个数增大,样本误差减小,修正效果得到改善,但增大到30以上时, 修正能力的提升就不再明显;观测数据标准差大小与RMSE值呈正相关;给定条件下当同化频率由1 h/次降低至2 h/次,整个模拟时长内的误差仍能得到较好控制,RMSE曲线并不会过快增长。     

Correlation of pmax and KSt to specific surface area and calorific value of a dust

The safety characteristics p_max and K_St of a flammable dust are usually determined in a closed apparatus (20 l- or 1 m³-vessel). The results are strongly influenced by the parameters of the apparatus (e.g. the volume, geometry, dispersion system). The aim of this project is to investigate whether there is a correlation between p_max and K_St on the one hand and the calorific value H_S and the specific surface area S_m of the dust on the other. In an experimental study, around 200 dust samples from different sectors of industry will be analysed. The analysis of around 102 dusts shows that the correlation may be described by a logarithmic function. Owing to differences in reaction mechanism it is reasonable to formulate different functions for different substance groups (e.g. metals, organic materials). This may lead to an algorithm which can predict p_max and K_St from H_S and S_m within a defined range of error.     

基于粗集理论算法-脊波神经网络的深基坑变形预测与应用

基坑开挖变形具有非线性特性,在脊波神经网络的基础上,采用粗集理论算法优化初始权值和阈值,建立了基于粗集理论算法-脊波神经网络的深基坑变形预测模型,应用该模型对西南地区某市火车站综合交通换乘中心南广场的基坑开挖过程进行了变形预测。结果表明:粗集理论算法能够对脊波神经网络进行优化,提高了脊波神经网络基坑变形预测结果的收敛速度和泛化能力;脊波神经网络能逼近基坑变形的非线性部分,避免了模型误差影响基坑开挖变形预测精度,提高了系统整体抗干扰性能。模型的预测值与实测值之间的误差在5%以内,满足实际工程的要求。     

Biomechanics of Brain and Spinal-Cord Injury: Analysis of Neuropathologic and Neurophysiology Experiments

This study analyzes 46 brain and 48 spinal-cord impact experiments. The velocity of brain impact was 2.0-10.0 m/s and displacement, 0.75-5.0 mm (5.3-33% compression) using a controlled pneumatic impact. The velocity of spinal-cord impact was 1.5-6.0 m/s and displacement, 1.25-3.25 mm (25-65% compression). Brain injury varied from cortical contusion, diffuse axonal injury (DAI), to fatalities, and spinal-cord injury from temporary to complete loss of somatosensory-evoked potentials. Logist functions were determined for each injury severity and various biomechanical parameters, VC, C, V, and combinations. Brain and spinal-cord injury is most strongly correlated to VC, the viscous response. The goodness-of-fit was x² = 22.1, R-0.84 and p< 0.0000 for fatal brain injury, x² = 27.5, R = 0.96 and p< 0.0000 for cortical contusion, and x² = 17.7, R = 0.49 and p < 0.0001 for partial recovery of spinal-cord conduction. Neural tissue is viscoelastic, with a rate-dependent tolerance related to energy absorption. VC is a measure of energy absorption by impact deformation and is predictive of neural contusion, DAI, long-duration coma, spinal-cord dysfunction, and death. Tolerances for various severities of neural injury are presented. At the tissue level, VC is the product of strain and strain-rate, ε dε/dt. The research shows that strain is not a sufficient parameter of neural injury risk, and that the product of strain and strain-rate is a key biomechanical parameter for brain and spinal-cord injury.     

基于循环神经网络的缓坡场地液化侧移预测*

为预测缓坡场地地震液化侧向位移,基于改进自适应算法（Rectified Adam）和循环神经网络模型（RNN）,提出液化侧移预测模型RA-RNN,通过对侧移数据进行样本学习,并利用改进自适应算法优化循环神经网络结构,验证RA-RNN模型可靠性,并与多元线性回归法（MLR）计算结果进行对比。结果表明:RA-RNN模型计算得到侧移一般为实测位移的0.7~1.3倍,训练结果R2,RMSE,MAE分别为0.977,0.375,0.141;土耳其科喀艾里RA-RNN模型预测结果RMSE和MAE为MLR模型的1/26,1/830;中国台湾集集镇RA-RNN模型预测结果RMSE和MAE为MLR模型的1/18,1/350,RA-RNN模型预测结果较优,预测精度及泛化能力得到很大提升。