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.

Dimensionality reduction and prediction of soil consolidation coefficient using random forest coupling with Relief algorithm

The consolidation coefficient of soil(C_(v))is a crucial parameter used for the design of structures leaned on soft soi.In general,the C_(v) is determined experimentally in the laboratory.However,the experimental tests are time-consuming as well as expensive.Therefore,researchers tried several ways to determine C_(v) via other simple soil parameters.In this study,we developed a hybrid model of Random Forest coupling with a Relief algorithm(RF-RL)to predict the C_(v) of soil.To conduct this study,a database of soil parameters collected from a case study region in Vietnam was used for modeling.The performance of the proposed models was assessed via statistical indicators,namely Coefficient of determination(R^(2)),Root Mean Squared Error(RMSE),and Mean Absolute Error(MAE).The proposal models were constructed with four sets of soil variables,including 6,7,8,and 13 inputs.The results revealed that all models performed well with a high performance(R^(2)>0.980).Although the RF-RL model with 13 variables has the highest prediction accuracy(R^(2)=0.9869),the difference compared with other models was negligible(i.e.,R^(2)=0.9824,0.9850,0.9825 for the cases with 6,7,8 inputs,respectively).Thus,it can be concluded that the hybrid model of RF-RL can be employed to predict C_(v) based on the basic soil parameters.

Optimization of machine learning models for predicting the compressive strength of fiber-reinforced self-compacting concrete

Fiber-reinforced self-compacting concrete(FRSCC)is a typical construction material,and its compressive strength(CS)is a critical mechanical property that must be adequately determined.In the machine learning(ML)approach to estimating the CS of FRSCC,the current research gaps include the limitations of samples in databases,the applicability constraints of models owing to limited mixture components,and the possibility of applying recently proposed models.This study developed different ML models for predicting the CS of FRSCC to address these limitations.Artificial neural network,random forest,and categorical gradient boosting(CatBoost)models were optimized to derive the best predictive model with the aid of a 10-fold cross-validation technique.A database of 381 samples was created,representing the most significant FRSCC dataset compared with previous studies,and it was used for model development.The findings indicated that CatBoost outperformed the other two models with excellent predictive abilities(root mean square error of 2.639 MPa,mean absolute error of 1.669 MPa,and coefficient of determination of 0.986 for the test dataset).Finally,a sensitivity analysis using a partial dependence plot was conducted to obtain a thorough understanding of the effect of each input variable on the predicted CS of FRSCC.The results showed that the cement content,testing age,and superplasticizer content are the most critical factors affecting the CS.

Using loose nanofiltration membrane for lake water treatment:A pilot study

Nanofiltration(NF)using loose membranes has a high application potential for advanced treatment of drinking water by selectively removing contaminants from the water,while membrane fouling remains one of the biggest problems of the process.This paper reported a seven-month pilot study of using a loose NF membrane to treat a sand filtration effluent which had a relatively high turbidity(∼0.4 NTU)and high concentrations of organic matter(up to 5 mg/L as TOC),hardness and sulfate.Results showed that the membrane demonstrated a high rejection of TOC(by<90%)and a moderately high rejection of two pesticides(54%–82%)while a moderate rejection of both calcium and magnesium(∼45%)and a low rejection of total dissolved solids(∼27%).The membrane elements suffered from severe membrane fouling,with the membrane permeance decreased by 70%after 85 days operation.The membrane fouling was dominated by organic fouling,while biological fouling was moderate.Inorganic fouling was mainly caused by deposition of aluminum-bearing substances.Though inorganic foulants were minor contents on membrane,their contribution to overall membrane fouling was substantial.Membrane fouling was not uniform on membrane.While contents of organic and inorganic foulants were the highest at the inlet and outlet region,respectively,the severity of membrane fouling increased from the inlet to the outlet region of membrane element with a difference higher than 30%.While alkaline cleaning was not effective in removing the membrane foulants,the use of ethylenediamine tetraacetate(EDTA)at alkaline conditions could effectively restore the membrane permeance.

Concentration levels of disinfection by-products in 14 swimming pools of China

Swimming has become a popular exercising and recreational activity in China but little is known about the disinfection by-products （DBPs） concentration levels in the pools. This study was conducted as a survey of the DBPs in China swimming pools, and to establish the correlations between the DBP concentrations and the pool water quality parameters. A total of 14 public indoor and outdoor pools in Beijing were included in the survey. Results showed that the median concentrations for total tfihalomethanes （TTHM）, nine haloacetic acids （HAA9）, chloral hydrate （CH）, four haloacetonitriles （HAN4）, 1,1- dichloropropanone, 1,1,1-trichloropropanone and trichlor- onitromethane were 33.8, 109.1, 30.1, 3.2, 0.3, 0.6 pg＇L-1 and below detection limit, respectively. The TTHM and HAA9 levels were in the same magnitude of that in many regions of the world. The levels of CH and nitrogenous DBPs were greatly higher than and were comparable to that in typical drinking water, respectively. Disinfection by chlorine dioxide or trichloroisocyanuric acid could sub- stantially lower the DBP levels. The outdoor pools had higher TTHM and HAA9 levels, but lower trihaloacetic acids （THAA） levels than the indoor pools. The TTHM and HAA9 concentrations could be moderately correlated with the free chlorine and total chlorine residuals but not with the total organic carbon （TOC） contents. When the DBP concentration levels from other survey studies were also included for statistical analysis, a good correlation could be established between the TTHM levels and the TOC concentration. The influence of chlorine residual on DBP levels could also be significant.

Impact of global climate change on stream low flows:A case study of the great Miami river watershed, Ohio, USA

Climate change will profoundly affect hydrological processes at various temporal and spatial scales.This study is focused on assessing the alteration of water resources availability and low flows frequencies driven by changing climates in different time periods of the 21st century.This study evaluates the adaptability of prevailing Global Circulation Models(GCMs)on a particular watershed through streamflow regimes.This analysis was conducted in the Great Miami River Watershed,Ohio by analyzing historical and future simulated streamflow using 10 climate model outputs and the Soil and Water Assessment Tool(SWAT).The climate change scenarios,consisting of ten downscaled Coupled Model Intercomparision Project Phase 5(CMIP5)climate models in combination with two Representative Concentration Pathways(RCP 4.5 and RCP 8.5)were selected based on the correlation between observed records and model outputs.Streamflow for three future periods,2016-2043,2044-2071 and 2072-2099,were independently analyzed and compared with the baseline period(1988-2015).Results from the average of ten models projected that 7-day low flows in the watershed would increase by 19%in the 21st century under both RCPs.This trend was also consistent for both hydrological(7Q10,1Q10)and biological low flow statistics(4B3,1B3).Similarly,average annual flow and monthly flows would also increase in future periods,especially in the summer.The flows simulated by SWAT in response to the majority of climate model projections showed a consistent increase in low flow patterns.However,the flow estimates using the Max-Planck-Institute Earth System Model(MPI-ESM-LR)climate output resulted in the biological based low flows(4B3,1B3)decreasing by 22.5%and 33.4%under RCP 4.5 and 56.9%and 63.7%under RCP 8.5,respectively,in the future when compared to the baseline period.Regardless,the low flow ensemble from the 10 climate models for the 21st century seemed to be slightly higher than that of historical low flows.

A hybrid machine learning model to estimate self-compacting concrete compressive strength

This study examined the feasibility of using the grey wolf optimizer(GWO)and artificial neural network(ANN)to predict the compressive strength(CS)of self-compacting concrete(SCC).The ANN-GWO model was created using 115 samples from different sources,taking into account nine key SCC factors.The validation of the proposed model was evaluated via six indices,including correlation coefficient(R),mean squared error,mean absolute error(MAE),IA,Slope,and mean absolute percentage error.In addition,the importance of the parameters affecting the CS of SCC was investigated utilizing partial dependence plots.The results proved that the proposed ANN-GWO algorithm is a reliable predictor for SCC’s CS.Following that,an examination of the parameters impacting the CS of SCC was provided.