Search

This paper represents the first study to use the grey model (GM) for impact prediction in the envi-ronmental impact assessment (EIA) of large com-munities. The relationship between the quantifiable impact and community scale factors of six communities was constructed. The impact of two others was predicted using GM (1, N) and multiple linear regression (MLR). The results indicate that when constructing a model, the mean absolute percentage errors (MAPEs) of GM (1, N) for assessed items lay between 0% and 16%, and the MAPEs of MLR lay between 1% and 17%, revealing good consistency; this considers topography/geology/soil, hydrology/water quality, air quality, noise and vibration, solid waste, fauna ecology, flora ecology, landscape aesthetic, land usage, social environment, and traffic. However, predictions for the MAPEs of the GM (1, N) lay between 3% and 19% in all but three items, whereas the MAPEs of the MLR lay between 1% and 20% except with two items. Because predicted community scale values were lower compared with those of other communities, the impact level was overestimated, resulting in a higher MAPE. GM (1, N) and MLR were able to predict the environmental impact and were used to analyze the likelihood of impact. In any future EIA review of a large community, the official committee could assess the likelihood of impact levels in such reports efficiently. [ABSTRACT FROM AUTHOR]

The operation quality of a wastewater treatment system influences its effluent quality, treatment cost, and performance stability. The aim of this research is to aid wastewater plants surpass extraordinary situations by using a neural network approach leading to an early warning system. Sensitivity analysis methods are proposed to evaluate influence and time interval. The factor majorly influencing the effluent quality is the recycle ratio, which has the fastest response time and can be selected as a crucial operational variable for the optimal dynamic operation of the wastewater treatment system. The results show that the use of statistical methods, network experiment designs, and sensitivity analyses of input or output variables of wastewater treatment plants can successfully build a wastewater prediction model. [ABSTRACT FROM AUTHOR]