Data-Driven Hazardous Gas Dispersion Modeling Using the Integration of Particle Filtering and Error Propagation Detection.

The accurate prediction of hazardous gas dispersion process is essential to air quality monitoring and the emergency management of contaminant gas leakage incidents in a chemical cluster. Conventional Gaussian-based dispersion models can seldom give accurate predictions due to inaccurate input parameters and the computational errors. In order to improve the prediction accuracy of a dispersion model, a data-driven air dispersion modeling method based on data assimilation is proposed by applying particle filter to Gaussian-based dispersion model. The core of the method is continually updating dispersion coefficients by assimilating observed data into the model during the calculation process. Another contribution of this paper is that error propagation detection rules are proposed to evaluate their effects since the measured and computational errors are inevitable. So environmental protection authorities can be informed to what extent the model output is of high confidence. To test the feasibility of our method, a numerical experiment utilizing the SF6 concentration data sampled from an Indianapolis field study is conducted. Results of accuracy analysis and error inspection imply that Gaussian dispersion models based on particle filtering and error propagation detection have better performance than traditional dispersion models in practice though sacrificing some computational efficiency.

Comparison of Machine Learning Models for Hazardous Gas Dispersion Prediction in Field Cases.

Dispersion prediction plays a significant role in the management and emergency response to hazardous gas emissions and accidental leaks. Compared with conventional atmospheric dispersion models, machine leaning (ML) models have both high accuracy and efficiency in terms of prediction, especially in field cases. However, selection of model type and the inputs of the ML model are still essential problems. To address this issue, two ML models (i.e., the back propagation (BP) network and support vector regression (SVR) with different input selections (i.e., original monitoring parameters and integrated Gaussian parameters) are proposed in this paper. To compare the performances of presented ML models in field cases, these models are evaluated using the Prairie Grass and Indianapolis field data sets. The influence of the training set scale on the performances of ML models is analyzed as well. Results demonstrate that the integrated Gaussian parameters indeed improve the prediction accuracy in the Prairie Grass case. However, they do not make much difference in the Indianapolis case due to their inadaptability to the complex terrain conditions. In addition, it can be summarized that the SVR shows better generalization ability with relatively small training sets, but tends to under-fit the training data. In contrast, the BP network has a stronger fitting ability, but sometimes suffers from an over-fitting problem. As a result, the model and input selection presented in this paper will be of great help to environmental and public health protection in real applications.