No abstract

Dust pollution in construction sites is an invisible hazard that is often ignored as a nuisance. Regulatory and engineering control methods are predominantly used for its mitigation. To control dust, dust-generating activities and their magnitudes need to be established. While researchers have comprehensively studied dust emissions of construction work, prediction of dust concentrations based on work phases and climatic conditions is still lacking. To overcome the above knowledge gap, this article selected two construction stages of a project to monitor dust generation using the HXF-35 dust sampler. Based on the collected data, dust emission characteristics of these two stages are studied, and dust emission characteristics under multiple pollution sources are analyzed. Based on the results, a BP neural network model is built to perform simulations of dust emission concentrations in different work areas and predict construction dust concentrations under different conditions. Except few, the majority of the work areas monitored have exceeded the allowable upper limit of TSP concentration stipulated by relevant standards. In addition, dust emission differences of work areas are pronounced. The results verified that the BP neural network dust concentration prediction model is feasible to be used to predict dust concentration changes in different work faces under different climate conditions and to provide a scientific base for pollution control. This study provides several practical solutions where the prediction of dust concentrations at designated work areas will allow construction companies early warning to implement mitigation measures before it becomes a serious health hazard. In addition, it provides an opportunity to re-evaluate those hazardous work in the light of these revelations. The outcome of this study is both original and useful for both construction companies and regulatory agencies. It can better predict the concentration of construction dust in different operating areas and different weather conditions and provide a guide for the prevention and control of construction dust.

The emotion-cause pair extraction task is a fine-grained task in text sentiment analysis, which aims to extract all emotions and their underlying causes in a document. Recent studies have addressed the emotion-cause pair extraction task in a step-by-step manner, i.e., the two subtasks of emotion extraction and cause extraction are completed first, followed by the pairing task of emotion-cause pairs. However, this fail to deal well with the potential relationship between the two subtasks and the extraction task of emotion-cause pairs. At the same time, the grammatical information contained in the document itself is ignored. To address the above issues, we propose a deep neural network based on span association prediction for the task of emotion-cause pair extraction, exploiting general grammatical conventions to span-encode sentences. We use the span association pairing method to obtain candidate emotion-cause pairs, and establish a multi-dimensional information interaction mechanism to screen candidate emotion-cause pairs. Experimental results on a quasi-baseline corpus show that our model can accurately extract potential emotion-cause pairs and outperform existing baselines.