Research on PM_(2.5) Concentration Prediction Algorithm Based on Temporal and Spatial Features

PM2.5 has a non-negligible impact on visibility and air quality as an important component of haze and can affect cloud formation and rainfall and thus change the climate,and it is an evaluation indicator of air pollution level.Achieving PM2.5 concentration prediction based on relevant historical data mining can effectively improve air pollution forecasting ability and guide air pollution prevention and control.The past methods neglected the impact caused by PM2.5 flow between cities when analyzing the impact of inter-city PM2.5 concentrations,making it difficult to further improve the prediction accuracy.However,factors including geographical information such as altitude and distance and meteorological information such as wind speed and wind direction affect the flow of PM2.5 between cities,leading to the change of PM2.5 concentration in cities.So a PM2.5 directed flow graph is constructed in this paper.Geographic and meteorological data is introduced into the graph structure to simulate the spatial PM2.5 flow transmission relationship between cities.The introduction of meteorological factors like wind direction depicts the unequal flow relationship of PM2.5 between cities.Based on this,a PM2.5 concentration prediction method integrating spatial-temporal factors is proposed in this paper.A spatial feature extraction method based on weight aggregation graph attention network(WGAT)is proposed to extract the spatial correlation features of PM2.5 in the flow graph,and a multi-step PM2.5 prediction method based on attention gate control loop unit(AGRU)is proposed.The PM2.5 concentration prediction model WGAT-AGRU with fused spatiotemporal features is constructed by combining the two methods to achieve multi-step PM2.5 concentration prediction.Finally,accuracy and validity experiments are conducted on the KnowAir dataset,and the results show that the WGAT-AGRU model proposed in the paper has good performance in terms of prediction accuracy and validates the effectiveness of the model.

A Hybrid Deep Learning Approach for PM2.5 Concentration Prediction in Smart Environmental Monitoring

Nowadays,air pollution is a big environmental problem in develop-ing countries.In this problem,particulate matter 2.5(PM2.5)in the air is an air pollutant.When its concentration in the air is high in developing countries like Vietnam,it will harm everyone’s health.Accurate prediction of PM2.5 concentrations can help to make the correct decision in protecting the health of the citizen.This study develops a hybrid deep learning approach named PM25-CBL model for PM2.5 concentration prediction in Ho Chi Minh City,Vietnam.Firstly,this study analyzes the effects of variables on PM2.5 concentrations in Air Quality HCMC dataset.Only variables that affect the results will be selected for PM2.5 concentration prediction.Secondly,an efficient PM25-CBL model that integrates a convolutional neural network(CNN)andBidirectionalLongShort-TermMemory(Bi-LSTM)isdeveloped.This model consists of three following modules:CNN,Bi-LSTM,and Fully connected modules.Finally,this study conducts the experiment to compare the performance of our approach and several state-of-the-art deep learning models for time series prediction such as LSTM,Bi-LSTM,the combination of CNN and LSTM(CNN-LSTM),and ARIMA.The empirical results confirm that PM25-CBL model outperforms other methods for Air Quality HCMC dataset in terms of several metrics including Mean Squared Error(MSE),Root Mean Squared Error(RMSE),Mean Absolute Error(MAE),and Mean Absolute Percentage Error(MAPE).

Deep-learning architecture for PM_(2.5) concentration prediction: A review

Accurately predicting the concentration of fine particulate matter(PM_(2.5))is crucial for evaluating air pollution levels and public exposure.Recent advancements have seen a significant rise in using deep learning(DL)models for forecasting PM_(2.5) concentrations.Nonetheless,there is a lack of unified and standardized frameworks for assessing the performance of DL-based PM_(2.5) prediction models.Here we extensively reviewed those DL-based hybrid models for forecasting PM_(2.5) levels according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses(PRISMA)guidelines.We examined the similarities and differences among various DL models in predicting PM_(2.5) by comparing their complexity and effectiveness.We categorized PM_(2.5) DL methodologies into seven types based on performance and application conditions,including four types of DL-based models and three types of hybrid learning models.Our research indicates that established deep learning architectures are commonly used and respected for their efficiency.However,many of these models often fall short in terms of innovation and interpretability.Conversely,models hybrid with traditional approaches,like deterministic and statistical models,exhibit high interpretability but compromise on accuracy and speed.Besides,hybrid DL models,representing the pinnacle of innovation among the studied models,encounter issues with interpretability.We introduce a novel three-dimensional evaluation framework,i.e.,Dataset-MethodExperiment Standard(DMES)to unify and standardize the evaluation for PM_(2.5) predictions using DL models.This review provides a framework for future evaluations of DL-based models,which could inspire researchers to standardize DL model usage in PM_(2.5) prediction and improve the quality of related studies.

Prediction of NO_(x)concentration using modular long short-term memory neural network for municipal solid waste incineration

Air pollution control poses a major problem in the implementation of municipal solid waste incineration(MSWI).Accurate prediction of nitrogen oxides(NO_(x))concentration plays an important role in efficient NO_(x)emission controlling.In this study,a modular long short-term memory(M-LSTM)network is developed to design an efficient prediction model for NO_(x)concentration.First,the fuzzy C means(FCM)algorithm is utilized to divide the task into several sub-tasks,aiming to realize the divide-and-conquer ability for complex task.Second,long short-term memory(LSTM)neural networks are applied to tackle corresponding sub-tasks,which can improve the prediction accuracy of the sub-networks.Third,a cooperative decision strategy is designed to guarantee the generalization performance during the testing or application stage.Finally,after being evaluated by a benchmark simulation,the proposed method is applied to a real MSWI process.And the experimental results demonstrate the considerable prediction ability of the M-LSTM network.

A prediction model of NH3 concentration for swine house in cold region based on Empirical Mode Decomposition and Elman neural network

In order to improve the accuracy and reliability of ammonia(NH3)concentration prediction,which can provides a support to the ventilation control strategy,so as to reduce the impact of NH3 on the health and productivity of swine,this paper proposed an NH3 concentration prediction method based on Empirical Mode Decomposition(EMD)and Elman neural network modelling.The NH3 concentration and other four environmental parameters including temperature,humidity,carbon dioxide and light intensity were decomposed into several different time-scale intrinsic mode functions(IMFs).Then,the Elman neural network prediction model was used to predict each IMF.The predicted NH3 was obtained by reconstructing all the IMFs by EMD.The results show that for the proposed method,the determination coefficient between the predicted and real measured value is 0.9856,the Mean Absolute Error is 0.7088 ppm,the Root Mean Square Error is 0.9096 ppm,and the Mean Absolute Percentage Error is 0.41%.Compared with the Elman neural network,the proposed method has a good improvement in the accuracy,and provide effective parameters for the environmental monitoring of the swine house and the regulation of the NH3 concentration.

Prediction of effluent concentration in a wastewater treatment plant using machine learning models

Of growing amount of food waste, the integrated food waste and waste water treatment was regarded as one of the efficient modeling method. However, the load of food waste to the conventional waste treatment process might lead to the high concentration of total nitrogen（T-N） impact on the effluent water quality. The objective of this study is to establish two machine learning models-artificial neural networks（ANNs） and support vector machines（SVMs）, in order to predict 1-day interval T-N concentration of effluent from a wastewater treatment plant in Ulsan, Korea. Daily water quality data and meteorological data were used and the performance of both models was evaluated in terms of the coefficient of determination（R^2）, Nash-Sutcliff efficiency（NSE）, relative efficiency criteria（d rel）. Additionally, Latin-Hypercube one-factor-at-a-time（LH-OAT） and a pattern search algorithm were applied to sensitivity analysis and model parameter optimization, respectively. Results showed that both models could be effectively applied to the 1-day interval prediction of T-N concentration of effluent. SVM model showed a higher prediction accuracy in the training stage and similar result in the validation stage.However, the sensitivity analysis demonstrated that the ANN model was a superior model for 1-day interval T-N concentration prediction in terms of the cause-and-effect relationship between T-N concentration and modeling input values to integrated food waste and waste water treatment. This study suggested the efficient and robust nonlinear time-series modeling method for an early prediction of the water quality of integrated food waste and waste water treatment process.

Modeling Tracer Flow Characteristics in Different Types of Pores: Visualization and Mathematical Modeling

Structure of porous media and fluid distribution in rocks can significantly affect the transport characteristics during the process of microscale tracer flow.To clarify the effect of micro heterogeneity on aqueous tracer transport,this paper demonstrates microscopic experiments at pore level and proposes an improved mathematical model for tracer transport.The visualization results show a faster tracer movement into movable water than it into bound water,and quicker occupancy in flowing pores than in storage pores caused by the difference of tracer velocity.Moreover,the proposed mathematical model includes the effects of bound water and flowing porosity by applying interstitial flow velocity expression.The new model also distinguishes flowing and storage pores,accounting for different tracer transport mechanisms(dispersion,diffusion and adsorption)in different types of pores.The resulting analytical solution better matches with tracer production data than the standard model.The residual sum of squares(RSS)from the new model is 0.0005,which is 100 times smaller than the RSS from the standard model.The sensitivity analysis indicates that the dispersion coefficient and flowing porosity shows a negative correlation with the tracer breakthrough time and the increasing slope,whereas the superficial velocity and bound water saturation show a positive correlation.

Short-term prediction of ammonia levels in goose houses via combined feature selector and random forest

Ammonia concentration(NH3)is a dominant source of environmental pollution in geese housing and profoundly affects the healthy growth of geese.Accurately forecasting NH3 and analyzing its change trends in geese houses is crucial for the survival of geese.A novel forecasting model was proposed by combining feature selector(CFS)and random forest(RF)to improve the prediction accuracy of NH3 in this study.The developed model integrated two modules.First,combining mutual information(MI)and relief-F,we propose that CFS quantify each feature’s importance values and eliminate the low-relation or unrelated features.Second,a random forest model was built using K-fold cross-validation grid search algorithm(CVGS)to obtain the RF hyperparameters to predict NH_(3).The simulation results show that the prediction accuracy was improved when feature selection after quantification based on the CFS was used.The mean square error(MSE),root mean square error(RMSE),and mean absolute percent error(MAPE)for the proposed model were 0.5072,0.6583,and 2.88%,respectively.The NH_(3) prediction model(CFS-CVGS-RF)based on Combined Feature Selector,cross-validation grid search algorithm(CVGS),and Random Forest(RF)exhibited the best prediction accuracy and generalization performance compared with other parallel forecasting models and is a suitable and useful tool for predicting NH3 in geese houses.The results of the research can provide a reference for the machine learning method to monitor the dynamic changes of ammonia in goose houses.

Predicted no-effect concentrations for mercury species and ecological risk assessment for mercury pollution in aquatic environment

Mercury（Hg） exists in different chemical forms presenting varied toxic potentials. It is necessary to explore an ecological risk assessment method for different mercury species in aquatic environment. The predicted no-effect concentrations（PNECs） for Hg（Ⅱ） and methyl mercury（Me Hg） in the aqueous phase, calculated using the species sensitivity distribution method and the assessment factor method, were 0.39 and 6.5 × 10-3μg/L, respectively. The partition theory of Hg between sediment and aqueous phases was considered, along with PNECs for the aqueous phase to conduct an ecological risk assessment for Hg in the sediment phase. Two case studies, one in China and one in the Western Black Sea, were conducted using these PNECs. The toxicity of mercury is heavily dependent on their forms,and their potential ecological risk should be respectively evaluated on the basis of mercury species.

To gel or not to gel：A prior prediction of gelation in solvent mixtures

The gelation behaviours of low molecular weight gelators 1,3：2,5：4.6-tris（3,4-dichlorobenzylidene）-Dmannitol（G1） and 2,4-（3.4-dichlorobenzylidene）-N-（3-aminopropyl）-D-gluconamide（G2） in 34 solvents have been studied.We found that sample dissolved at low concentrations may become a gel or precipitate at higher concentrations.The Hansen solubility parameters（HSPs） and a Teas plot were employed to correlate the gelation behaviours with solvent properties,but with no success if the concentration of the tests was not maintained constant.Instead,on the basis of the gelation results obtained for the G1 and G2 in single solvents,we studied the gelation behaviours of G1 and G2 in23 solvent mixtures and found that the tendency of a gelator to form a gel in mixed solvents is strongly correlated with its gelation behaviours in good solvents.If the gelation occurs in a good solvent at higher concentrations,it will take place as well in a mixed solvent（the good solvent plus a poor solvent） at a certain volume ratio.In contrast,if the gelator forms a precipitate in a good solvent at higher concentrations,no gelation is to be observed in the mixed solvents.A gelation rule for mixed solvents is thus proposed,which may facilitate decision making with regard to solvent selection for gel formation in the solvent mixtures in practical applications.

The potential risk assessment for different arsenic species in the aquatic environment

The different toxicity characteristics of arsenic species result in discrepant ecological risk.The predicted no-effect concentrations（PNECs） 43.65, 250.18, and 2.00 × 10^3μg/L were calculated for As（III）, As（V）, and dimethylarsinic acid in aqueous phase, respectively. With these PNECs, the ecological risk from arsenic species in Pearl River Delta in China and Kwabrafo stream in Ghana was evaluated. It was found that the risk from As（III） and As（V）in the samples from Pearl River Delta was low, while much high in Kwabrafo stream. This study implies that ecological risk of arsenic should be evaluated basing on its species.