Application of machine learning technique for predicting and evaluating chloride ingress in concrete

The degradation of concrete structure in the marine environment is often related to chloride-induced corrosion of reinforcement steel. Therefore, the chloride concentration in concrete is a vital parameter for estimating the corrosion level of reinforcement steel. This research aims at predicting the chloride content in concrete using three hybrid models of gradient boosting (GB), artificial neural network (ANN), and random forest (RF) in combination with particle swarm optimization (PSO). The input variables for modeling include exposure condition, water/binder ratio (W/B), cement content, silica fume, time exposure, and depth of measurement. The results indicate that three models performed well with high accuracy of prediction (R²≥ 0.90). Among three hybrid models, the model using GB_PSO achieved the highest prediction accuracy (R² = 0.9551, RMSE = 0.0327, and MAE = 0.0181). Based on the results of sensitivity analysis using SHapley Additive exPlanation (SHAP) and partial dependence plots 1D (PDP-1D), it was found that the exposure condition and depth of measurement were the two most vital variables affecting the prediction of chloride content. When the number of different exposure conditions is larger than two, the exposure significantly impacted the chloride content of concrete because the chloride ion ingress is affected by both chemical and physical processes. This study provides an insight into the evaluation and prediction of the chloride content of concrete in the marine environment.     

Novel hybrid models of ANFIS and metaheuristic optimizations (SCE and ABC) for prediction of compressive strength of concrete using rebound hammer field test

In this study, we developed novel hybrid models namely Adaptive Neuro Fuzzy Inference System (ANFIS) optimized by Shuffled Complex Evolution (SCE) on the one hand and ANFIS with Artificial Bee Colony (ABC) on the other hand. These were used to predict compressive strength (Cs) of concrete relating to thirteen concrete-strength affecting parameters which are easy to determine in the laboratory. Field and laboratory tests data of 108 structural elements of 18 concrete bridges of the Ha Long-Van Don Expressway, Vietnam were considered. The dataset was randomly divided into a 70:30 ratio, for training (70%) and testing (30%) of the hybrid models. Performance of the developed fuzzy metaheuristic models was evaluated using standard statistical metrics: Correlation Coefficient (R), Root Mean Square Error (RMSE) and Mean Absolute Error (MAE). The results showed that both of the novel models depict close agreement between experimental and predicted results. However, the ANFIS-ABC model reflected better convergence of the results and better performance compared to that of ANFIS-SCE in the prediction of the concrete Cs. Thus, the ANFIS-ABC model can be used for the quick and accurate estimation of compressive strength of concrete based on easily determined parameters for the design of civil engineering structures including bridges.     

Assessment of different machine learning techniques in predicting the compressive strength of self-compacting concrete

The compressive strength of self-compacting concrete (SCC) needs to be determined during the construction design process. This paper shows that the compressive strength of SCC (CS of SCC) can be successfully predicted from mix design and curing age by a machine learning (ML) technique named the Extreme Gradient Boosting (XGB) algorithm, including non-hybrid and hybrid models. Nine ML techniques, such as Linear regression (LR), K-Nearest Neighbors (KNN), Support Vector Machine (SVM), Decision Trees (DTR), Random Forest (RF), Gradient Boosting (GB), and Artificial Neural Network using two training algorithms LBFGS and SGD (denoted as ANN_LBFGS and ANN_SGD), are also compared with the XGB model. Moreover, the hybrid models of eight ML techniques and Particle Swarm Optimization (PSO) are constructed to highlight the reliability and accuracy of SCC compressive strength prediction by the XGB_PSO hybrid model. The highest number of SCC samples available in the literature is collected for building the ML techniques. Compared with previously published works’ performance, the proposed XGB method, both hybrid and non-hybrid models, is the most reliable and robust of the examined techniques, and is more accurate than existing ML methods (R² = 0.9644, RMSE = 4.7801, and MAE = 3.4832). Therefore, the XGB model can be used as a practical tool for engineers in predicting the CS of SCC.     

Uptake and release of zinc by aquatic bryophytes (Fontinalis antipyretica L. ex. Hedw.)

The zinc uptake and posterior release by an aquatic bryophyte—Fontinalis antipyretica L. Ex Hedw.—was experimentally studied in laboratory exposing the plants to different zinc concentrations in the range, 1.0–5.0 mg l⁻¹, for a 144 h contamination period, and then exposed to metal-free water for a 120 h decontamination period. The experiments were carried out in perfectly mixed contactors at controlled illumination, using mosses picked out in February 1997, with a background initial zinc concentration of 263 mg g⁻¹ (dry wt.). A first-order mass transfer kinetic model was fitted to the experimental data to determine the uptake and release constants, k₁ and k₂, the zinc concentration in mosses at the end of the uptake period, C_mu, and at the equilibrium, for the contamination and decontamination stages, C_me and C_mr, respectively. A bioconcentration factor, BCF=k₁/k₂ (zinc concentration in the plant, dry wt./zinc concentration in the water) was determined. A biological elimination factor defined as BEF=1−C_mr/C_mu was also calculated. BCF decreases from about 4500 to 2950 as Zn concentration in water increases from 1.05 to 3.80 mg l⁻¹. BEF is approximately constant and equal to 0.80. Comparing Zn and Cu accumulation by Fontinalis antipyretica, it was concluded that the uptake rate for Zn (145 h⁻¹) is much lower than for Cu (628 h⁻¹) and the amount retained by the plant decreased by a factor of about seven.     

Pavement sustainability index for highway infrastructures: A case study of Maryland

Pavement deterioration creates conditions that undermine their performances, which gives rise to the need for maintenance and rehabilitation. This paper develops a mathematical multi-linear regression analysis (MLRA) model to determine a pavement sustainability index (PSTI) as dependent variable for flexible pavements in Maryland. Four categories of pavement performance evaluation indicators are subdivided into seven pavement condition indices and analyzed as independent variables for each section of pavement. Data are collected from five different roadways using field evaluations and existing database. Results indicate that coefficient of determination (R²) is correlated and significant, R² = 0.959. Of the seven independent variables, present serviceability index (PSI) is the most significant with a coefficient value of 0.032, present serviceability rating (PSR) coefficient value= 0.028, and international roughness index (IRI) coefficient value= ?0.001. Increasing each unit value of coefficients for PSI and PSR would increase the value of PSTI; thereby providing a more sustainable pavement infrastructure; which explains the significance of the model and why IRI will most likely impact environmental, economic and social values.     

Toxicity of chlorine to juvenile spot, Leiostomus Xanthurus

The sensitivity of juvenile spot, Leiostomus xanthurus, to total residual chlorine (TRC) in flowing sea-water was investigated. Incipient LC₅₀ bioassays, histopathology, avoidance tests and the combined effect of thermal stress and TRC were used to assess sensitivity.

Estimated incipient LC₅₀ values were 0.12 mg 1⁻¹ TRC at 10°C and 0.06 mg 1⁻¹ TRC at 15°C. Histological examination of spot used in the incipient LC₅₀ bioassay at 15°C and sacrificed while alive indicated pseudobranch and gill damage occurred in individuals exposed to a measured TRC concentration of 1.57 mg 1⁻¹. Spot exposed to lower concentrations of TRC, 0.02 0.06 mg 1⁻¹ at 15°C and sacrificed alive showed no consistent tissue damage.

Spot demonstrated temperature dependent avoidance responses to TRC. At 10°C, a concentration of 0.18 mg 1⁻¹ was required for significant (X²; P < 0.05) avoidance; at 15 and 20°C, spot showed significant avoidance of TRC concentrations as low as 0.05 mg 1⁻¹.

Simultaneous exposure of spot to thermal stress (5, 10 or 13°C above the acclimation temperature of 15°C) at measured TRC concentrations of 0.05 0.07 and 0.34–0.52 mg 1⁻¹ demonstrated a significant, (Z²) with Yates correction, P < 0.05) increase in sensitivity to TRC with increased temperature and exposure times for some of the groups tested.     

Fine-spray (water mist) protection of shipboard engine rooms

Twenty-three fire tests were conducted to determine the ability of current fine-spray (water mist) technologies to extinguish fires specified in the International Maritime Organization (IMO) fire test procedure for engine rooms greater than 3000 m³. The tests were conducted using nozzle installed at a 5 m height and 1·5 m spacing in a large test facility (2800 m² area and 18 m height). Two types of nozzles were tested: a low pressure nozzle operating between 1·2 and 1·5 MPa with flow per nozzle between 12·0 and 13·41/m and a multi-nozzle high-pressure prototype consisting of seven nozzles operating at 6·9 MPa flowing at 5·3 1/m per prototype. These nozzles were selected because they had been shown to extinguish IMO fire tests in enclosures with a protected area of 83 m² and a ceiling height of 4·5 m. The fire tests selected from the IMO procedure included 6 MW diesel spray fires on top of the IMO engine mock-up, a shielded 6 MW diesel spray fire adjacent to the mock-up, a 1 MW shielded diesel spray fire adjacent to the mock-up, and a wood crib within a 2 m² pan filled with heptane. In tests in which no additional enclosure surrounded the nozzles other than the test facility, fires were not significantly affected by the water mist using either nozzle. To further investigate mist-system capabilities, a ceiling was then placed directly over the nozzles at a 5 m height covering an area of 188 m². Using 90 high-pressure prototype nozzles, the test fires were not extinguished. A 940 m³ enclosure was then formed by dropping tarpaulins to the floor from the ceiling. A 4 m² vent was placed in the wall. With the 90 high-pressure prototype nozzles, the 6 MW spray fire on top of the mock-up was extinguished. When the 6 MW fire was shielded beside the mock-up, the fire was not extinguished. With the vent closed, the 6 MW shielded spray fire was extinguished. Under the same test conditions, a 1 MW shielded diesel spray fire and a 0·1 m² heptane pool fire were not extinguished. The fire test results indicated that protection of engine rooms with volumes of about 1000 m³ is possible by optimizing current fine-spray technology while significantly larger volumes will require improved discharge characteristics.     

Influence of inter-annual variations in climatic factors on fecal coliform levels in Mississippi Sound

Information on the effects of inter-annual variations in climatic factors on fecal coliform levels in coastal waters is scarce. We used 11 years (1990–2001) of rainfall, water temperature, Pearl River stage and salinity data to assess the effects of the 1991–1992 and 1997–1998 El Niño events on fecal coliform levels in Mississippi Sound. El Niño-Southern Oscillation had major effects on Pearl River stage and water quality in the Sound. The geometric mean fecal coliform number differed among years (P=0.0001), being highest during 1991–1992 El Niño years (14.22 MPN per 100 ml) and lowest during 1999–2000 La Niña years (<1.8 MPN per 100 ml). Mean salinity varied among years (P=0.0001) from 9 ppt (1991–1992) to 21 ppt (1999–2000). Mean water temperature was lowest in 1997–1998 (14.5 °C) and highest in 1998–2000 (19.4 °C). Pearl River stage differed among years (P=0.0001), ranging from 1.96 m (1999–2000) to 3.57 m (1997–1998). Inverse relationships were observed between fecal coliform levels and salinity (r²=0.74; P=0.001) and water temperature (r²=0.69; P=0.001), whereas positive relationships were obtained with total rainfall (r²=0.52; P=0.013) and Pearl River stage (r²=0.90; P=0.0001). These relationships are useful for evaluating the potential effects of climate change on water quality and classification of shellfish growing waters in order to protect humans from consuming contaminated shellfish in shallow river-influenced estuaries.     

Presentation of regression analysis,GP and GMDH models to predict the pedestrian density in various urban facilities

In this study, the relationship between space mean speed (SMS), flow rate and density of pedestrians was investigated in different pedestrian facilities, including 1 walkway, 2 sidewalks, 2 signalized crosswalks and 2 mid-block crosswalks. First, statistical analysis was performed to investigate the normality of data and correlation of variables. Regression analysis was then applied to determine the relationship between SMS, flow rate, and density of pedestrians. Finally, two prediction models of density were obtained using genetic programming (GP) and group method of data handling (GMDH) models, and k-fold and holdout cross-validation methods were used to evaluate the models. By the use of regression analysis, the mathematical relationships between variables in all facilities were calculated and plotted, and the best relationships were observed in flow rate-density diagrams. Results also indicated that GP had a higher R² than GMDH in the prediction of pedestrian density in terms of flow rate and SMS, suggesting that GP was better able to model SMS and pedestrian density. Moreover, the application of k-fold cross-validation method in the models led to better performances compared to the holdout cross-validation method, which shows that the prediction models using k-fold were more reliable. Finally, density relationships in all facilities were obtained in terms of SMS and flow rate.     

理想砂井地基径向弹黏塑性固结分析

饱和黏土变形的非线性和流变特性已被人们所逐渐认识。为进一步探讨该特性对砂井地基固结过程的影响,引入考虑时间效应的统一硬化(UH)本构模型描述土体骨架弹黏塑性变形关系,在Barron砂井固结理论的基础上,重新推导砂井径向固结方程,并给出方程的隐式有限差分格式。通过与Berry基于压缩指数的显式数值解对比,验证本文计算方法的有效性。探讨了UH模型参数和渗透指数对砂井径向固结的影响。数值结果表明:黏性土的粘滞性使得固结前期砂井地基内部出现了孔压升高现象,并延缓了固结中后期砂井地基内孔隙水压力的整体消散。同时,随着初始超固结参数R₀、回弹指数C_s的增大或渗透指数C_k的减小,此现象变得更加明显。     

竖向受压支盘桩沉降非线性简化分析方法

为研究支盘桩沉降及相关力学性状,考虑到桩土接触非线性,针对支盘桩,在桩侧、桩端及扩径体端部均采用双曲线型荷载传递函数描述其力学行为,应用分段位移协调迭代算法,提出了一种计算支盘桩沉降及承载性状的非线性简化分析方法。与硬塑性黏土地基和淤泥质软土地基的两种支盘桩现场试验作对比,得到的计算值与实测值有较高的吻合度,证明了方法的合理性和适用性。通过参数分析,发现桩土摩擦角δ及水平土压力系数比k/k₀对承载力有较大影响,且承载力随着二者增大而增大;破坏比R_f及端承土体破坏角θ对承载力影响偏小,承载力随着θ增大而增大,随着R_f的增大而减小。     

Probabilistic approach to elastic-plastic coupled instability

Interaction of elastic and plastic instability modes is considered. Buckling resistance of columns is treated as the lesser value of plastic resistance R_pl of a critical cross-section and the Euler critical force R_cr. The random variables R_pl and R_cr have log-normal probability distributions. Their coefficients of variation are supposed equal and independent from the slenderness ratio λ in the first model of random instability. A simple formula for the buckling factor is derived thanks to the introduction of equivalent Weibull distribution functions. Strict formulae are derived in the second model of coupled instability for the columns of equal median values R_pl = R_cr when the relative slenderness ratthe columns λ( = 1. The design values R_d of the elastic-plastic resistance are reduced by an analysis factor C which depends on the slenderness λ( and the imperfection class, a, b, c or d, defined for European multiple buckling curves. Optimal values of resistance factor γ_R and analysis factor γ_C are determined so that the risks of exceeding the design values R_d and C_d are equal.     

溶蚀孔隙型灰岩岩体宏观力学参数取值研究

灰岩地层具有较好的可靠性和适宜性,并具有较大的各向异性,其分布从大的溶洞、管道、断裂到小的溶蚀裂隙、孔洞等,前者在工程探测及处理措施方面研究已较为成熟,笔者针对溶蚀孔、隙型灰岩的力学参数取值进行了研究,得到了如下结论:溶蚀孔、隙型灰岩岩石抗压强度与孔隙率存在较好的幂函数关系,幂常量接近于-0.5,即y=ax^-0.5,其系数a代表了孔隙率为1%时的岩石抗压强度;基于Hoek-Brown强度准则及岩体孔隙率n,建立岩体参数m_b、S、a与孔隙率n的关系式,并通过孔隙率来描述岩体力学参数(岩体抗压强度R_cm、岩体抗拉强度R_tm、岩体抗剪强度参数c_m和φ_m),简化了通过量化GSI进行力学参数取值的复杂性,避免了量化GSI的主观性和不确定性,在溶蚀孔、隙型灰岩地层具有较好的可靠性和适宜性,对同类场区的工程建设具有一定的指导意义。     

循环剪切作用下泥质软岩本构模型研究

库区微震频发现象使泥质软岩受循环剪切动力作用。基于分数阶西原模型,将Able黏壶进行改进,获得相应条件下的本构模型,并通过泥质软岩试样循环剪切试验对该本构模型进行了验证。研究表明:①原有分数阶西原本构模型可归纳为本模型正弦函数(循环作用)为零时的特殊解;②模型表现出较好的适用性(R²∈(0.96,0.98)),且方程曲线具备试验曲线的“单双拐点”特征;③模型主要影响因素为分数阶阶数n,其中稳定型方程n∈(0,0.6)、破坏型方程n∈(0.6,1.0);④随n的增加,抗剪模量G与黏滞系数η大致呈现降低趋势,可较好反映宏观现象;⑤从参数扰动的角度解释了以往文献出现的η随n增加可能表现出的无规律波动现象。研究成果可为微震条件下含泥质软岩边坡长期稳定性控制提供理论参考。     

Algal growth in warm temperate reservoirs: kinetic examination of nitrogen, temperature, light, and other nutrients

Nutrient limited growth of the phytoplankton assemblage in two Texas reservoirs was studied by a combination of nutrient addition experiments and statistical modeling. Dilution bioassays were run to ascertain the qualitative and quantitative patterns in nutrient limitation. Algal growth was frequently and strongly nutrient limited, particularly when temperature was >22°C. By itself, N was more often stimulatory than P, though strong additional enhancement of growth by P and trace nutrients was often detected. Monod growth kinetics indicated that half-saturation constants for N limited growth for the entire algal assemblage were in the range 20–200 μg N/L, relatively high compared to literature values, and increased with increasing temperature. Maximal growth was also an increasing function of temperature. A single temperature-dependent model was fit to the growth dynamics for all experiments showing N-limitation. The model μ=0.0256·T([DIN]/66.0+[DIN]) where μ is specific growth rate (d⁻¹), T is temperature (°C) and [DIN] is dissolved inorganic N (μmol/L) fit the experimental results reasonably well (r²=0.82). However, only a modest predictive power for growth in the controls (our best estimate of growth in situ) was achieved (r²=0.26). Thus, even with unusually detailed, site-specific fitting of model parameters, accurately modeling algal growth in natural ecosystems can remain a challenge.     

Kinetics of phenol degradation by activated sludge: Value of measurements in a batch reactor

Kinetic measurements on the degradation of phenol in a batch reactor were done at 20°C using an activated sludge taken from a continuous reactor where the process had shown Monod-type kinetics without substrate inhibition. When employing a sludge sample, adapted to the substrate steady concentration C_s¹, for runs with initial substrate concentration C_s⁰, observed batch rates were as high as expected for C_s⁰ = C_s¹, but lower than expected in runs carried out at C_s⁰ > C_s1¹, with indications of a transient inhibitory effect instead of a positive response to the increased substrate concentration. It is concluded that kinetic data collected in a batch reactor are of limited value for the design of a continuous biodegradation process.     

A contribution to environmental risk assessment for transport of cadmium through groundwater layers. Case study of the Sava river (near Zagreb, Croatia) region

Adsorption capacities of cadmium on aquifer material of the Sava River alluvial sediment were determined as a function of flow rate, pH and presence of organic coating using laboratory column technique. In a permeameter with constant hydrostatic pressure, a laboratory coefficient of permeability, K_l, and a specific coefficient of permeability, K_s, have been determined. The value of 28.6 Darcy for specific permeability shows that the sediment belongs to a good aquifer (permeability >1 Darcy). The adsorption capacity of cadmium at the bottom of the breakthrough curve, C_sb, of a fresh sediment at pH 5.8 varied from 0.40 to 0.45 mg g⁻¹ at axial flow rate between 98 and 501 cm h⁻¹. Capacity values of maximum adsorption, C_s, were in the range from 0.67 to 0.72 mg g⁻¹. This implies a significance of C_sb values in risk assessment studies concerning a discharge of cadmium into rivers and lakes and its input to groundwater layers. Distribution coefficients, K_d, were between 26.3 and 28.2 ml g⁻¹. In order to create a more reproducible column, a fraction between 125 and 250 μm was used. In that case C_s values varied from 0.25 mg g⁻¹ at pH 2.5 (organic coating present) to 1.12 mg g⁻¹ at pH 5.8 (organic coating removed).     

Degradation of polychlorinated dibenzo-p-dioxin and dibenzofuran contaminants in 2,4,5-T by photoassisted iron-catalyzed hydrogen peroxide

The objective of this study was to determine the fate of polychlorinated dibenzo-p-dioxin (PCDD) and dibenzofuran (PCDF) contaminants in the herbicide 2,4,5-trichlorophenoxyacetic acid after treatment by photoassisted Fe³⁺-catalyzed hydrogen peroxide oxidation, which is being investigated for waste treatment. Toluene-extraction of aqueous reagent grade 2,4,5-T gave 30 quantifiable PCDD/F isomer peaks, some of which represented more than one congener. They included 13 tetra- through heptachloro PCDDs totalling 18.4 ng/g, and 17 tetra- through octachloro PCDFs totalling 12.9 ng/g. Treatment with Fe³⁺/H₂O₂/hv (mole ratio 2,4,5-T:Fe³⁺:H₂O₂ = 1:10:200; visible fluorescent light, 300 nm cutoff) in aerated solution at pH 2.8 resulted in 89–100% removal of all PCDD/F peaks except that for OCDF, which was 66% removed. The 2,4,5-T was shown previously to be completely mineralised to HCl and CO₂. Direct photolysis of the PCDD/Fs was negligible. The 2,3,7,8-tetra-CDD referenced “toxic equivalent” used in risk assessment was reduced by 99.9% (U.S. Environmental Protection Agency method) or 99.6% (international method). Reaction mixtures supplemented with 2,3,7,8-(UL-¹⁴C)-tetra-CDD and 2,3,4,7,8-(dichlorophenyl-¹⁴C)-penta-CDF evolved up to 45 and 64%, respectively, theoretical yield of ¹⁴CO₂. Most of the remaining ¹⁴C was converted to hydrophilic forms not extractable by toluene. Dark (Fe³⁺/H₂O₂) treatment was somewhat less effective overall.     

Development of deep neural network model to predict the compressive strength of FRCM confined columns

The present study describes a reliability analysis of the strength model for predicting concrete columns confinement influence with Fabric-Reinforced Cementitious Matrix (FRCM). through both physical models and Deep Neural Network model (artificial neural network (ANN) with double and triple hidden layers). The database of 330 samples collected for the training model contains many important parameters, i.e., section type (circle or square), corner radius r_c, unconfined concrete strength f_co, thickness n_t, the elastic modulus of fiber E_f , the elastic modulus of mortar E_m. The results revealed that the proposed ANN models well predicted the compressive strength of FRCM with high prediction accuracy. The ANN model with double hidden layers (APDL-1) was shown to be the best to predict the compressive strength of FRCM confined columns compared with the ACI design code and five physical models. Furthermore, the results also reveal that the unconfined compressive strength of concrete, type of fiber mesh for FRCM, type of section, and the corner radius ratio, are the most significant input variables in the efficiency of FRCM confinement prediction. The performance of the proposed ANN models (including double and triple hidden layers) had high precision with R higher than 0.93 and RMSE smaller than 0.13, as compared with other models from the literature available.     

The energy balance of an exterior window surface, Inuvik, N.W.T., Canada

This paper examines the energy loss of an exterior window surface during the month of January, 1975, in Inuvik, N.W.T., Canada. Radiative exchanges (L^*) were measured and the energy loss through the window was calculated (Q_F). Since an energy balance exist at the exterior window surface, a heat transfer coefficient (h) was determined by residual: h=(Q_F−L^*)/(T_s−T_a). Calculated values for the heat transfer coefficient compare favourably with those of other studies. The atmospheric controls (i.e. wind speed and cloud cover) on radiative and convective heat losses and their influence on urban climate are discussed with respect to the observational data.