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

2018 ◽  
Vol 15 (8) ◽  
pp. 1640 ◽  
Author(s):  
Zhengqiu Zhu ◽  
Sihang Qiu ◽  
Bin Chen ◽  
Rongxiao Wang ◽  
Xiaogang Qiu
Keyword(s):  
Particle Filtering ◽  
Error Propagation ◽  
Dispersion Model ◽  
Data Driven ◽  
Dispersion Modeling ◽  
Dispersion Models ◽  
Concentration Data ◽  
Gas Dispersion ◽  
Computational Errors ◽  
Hazardous Gas

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.

Toxic gas dispersion models: Can they predict protective action distances in case of a chemical spill?

2008 ◽  
Vol 6 (5) ◽  
pp. 23
Author(s):  
John S. Nordin, PhD
Keyword(s):  
Dispersion Model ◽  
Protective Action ◽  
Weather Conditions ◽  
Dispersion Modeling ◽  
Dispersion Models ◽  
Gas Dispersion ◽  
The Public ◽  
Emergency Responders ◽  
Toxic Concentration ◽  
Stable Conditions

Emergency responders often use a gas dispersion model to estimate downwind airborne concentrations of a toxic chemical in case of a chemical spill accident. For protecting the public, a protective action distance from the spill source is established based on the distance where the toxic concentration drops below some level of concern. This distance is used as a basis for evacuation of the public from the area or for instructions to shelter-in-place. However, in real-world accidents, the responders neither know the amount of chemicals released into the air nor the duration of the release, and moreover, the concentrations of chemicals at any location will vary over time. Depending on what input information is put into the model, different results will be obtained. The problem of what input parameters to use for gas dispersion modeling is illustrated for a hypothetical 90-ton chlorine railcar accident, where the railcar is breached. Different answers for a protective action distance are obtained depending on whether the tables in the Emergency Response Guidebook or any of the popular gas dispersion models are used. Very different answers are obtained from any model depending on whether whole of the chemical is released at once as a gas or aerosol or whether the liquefied chlorine evaporates slowly inside a ruptured 90-ton railcar tank, and also the weather conditions. To avoid misunderstandings, people who use models to establish a protective action distance must also communicate the circumstances in which the models are used, eg, “worst possible what-if scenario,” etc, or “nighttime stable conditions,” or other situations.

scholarly journals Intercomparison of atmospheric trace gas dispersion models: Barnett Shale case study

2019 ◽  
Vol 19 (4) ◽  
pp. 2561-2576 ◽  
Author(s):  
Anna Karion ◽  
Thomas Lauvaux ◽  
Israel Lopez Coto ◽  
Colm Sweeney ◽  
Kimberly Mueller ◽  
...  
Keyword(s):  
Dispersion Model ◽  
Methane Emissions ◽  
Gas Production ◽  
Estimation Methods ◽  
Trace Gas ◽  
Barnett Shale ◽  
Dispersion Models ◽  
Data Set ◽  
Gas Dispersion ◽  
Tracer Dispersion

Abstract. Greenhouse gas emissions mitigation requires understanding the dominant processes controlling fluxes of these trace gases at increasingly finer spatial and temporal scales. Trace gas fluxes can be estimated using a variety of approaches that translate observed atmospheric species mole fractions into fluxes or emission rates, often identifying the spatial and temporal characteristics of the emission sources as well. Meteorological models are commonly combined with tracer dispersion models to estimate fluxes using an inverse approach that optimizes emissions to best fit the trace gas mole fraction observations. One way to evaluate the accuracy of atmospheric flux estimation methods is to compare results from independent methods, including approaches in which different meteorological and tracer dispersion models are used. In this work, we use a rich data set of atmospheric methane observations collected during an intensive airborne campaign to compare different methane emissions estimates from the Barnett Shale oil and natural gas production basin in Texas, USA. We estimate emissions based on a variety of different meteorological and dispersion models. Previous estimates of methane emissions from this region relied on a simple model (a mass balance analysis) as well as on ground-based measurements and statistical data analysis (an inventory). We find that in addition to meteorological model choice, the choice of tracer dispersion model also has a significant impact on the predicted downwind methane concentrations given the same emissions field. The dispersion models tested often underpredicted the observed methane enhancements with significant variability (up to a factor of 3) between different models and between different days. We examine possible causes for this result and find that the models differ in their simulation of vertical dispersion, indicating that additional work is needed to evaluate and improve vertical mixing in the tracer dispersion models commonly used in regional trace gas flux inversions.

scholarly journals Diagnosing the Surface Layer Parameters for Dispersion Models within the Meteorological-to-Dispersion Modeling Interface

2010 ◽  
Vol 49 (2) ◽  
pp. 221-233 ◽  
Author(s):  
M. Sofiev ◽  
E. Genikhovich ◽  
P. Keronen ◽  
T. Vesala
Keyword(s):  
Surface Layer ◽  
Weather Forecasting ◽  
Dispersion Model ◽  
Temporal Correlation ◽  
Complex Surface ◽  
Dispersion Modeling ◽  
Limited Area ◽  
Dispersion Models ◽  
Absolute Flux ◽  
The Absolute

Abstract The problem of providing dispersion models with meteorological information from general atmospheric models used, for example, for weather forecasting is considered. As part of a generalized meteorological-to-dispersion model interface, a noniterative scheme diagnosing the surface layer characteristics from wind, temperature, and humidity profiles was developed. The scheme verification included long-term comparison with data of meteorological masts at Cabauw, the Netherlands, and Hyytiälä, Finland. The algorithm compatibility and consistency with the High-Resolution Limited-Area Model (HIRLAM) was also checked, as this model is routinely used as a meteorological driver for the Air Quality and Emergency Modeling System (SILAM). The comparison with Cabauw mast data showed a good quantitative agreement between observed and diagnosed heat and momentum fluxes: the temporal correlation coefficient was ∼0.8, bias was less than 10% of the absolute flux levels, regression slope deviated from unity for less than 20% with the intercept being less than 10% of the absolute flux values, and so on. In the case of complex surface features (Hyytiälä mast in forest) the scheme proved to be robust with large deviations appearing only if the input profile data were taken outside the constant-flux layer. Comparison with the HIRLAM model showed qualitatively good agreement but also highlighted several differences between the goals, standards, and methodologies of meteorological and dispersion models. The scheme was implemented in SILAM, which served as the development platform.

scholarly journals GEOSTATISTICS FOR AIR QUALITY MAPPING: CASE OF BAGUIO CITY, PHILIPPINES

2019 ◽  
Vol XLII-4/W19 ◽  
pp. 353-359
Author(s):  
R. V. Ramos ◽  
A. C. Blanco
Keyword(s):  
Air Quality ◽  
Dispersion Model ◽  
Atmospheric Dispersion ◽  
Spatial Variations ◽  
Dispersion Modeling ◽  
Dispersion Models ◽  
Night Time ◽  
Smoothing Factor ◽  
Pollutant Concentrations ◽  
Atmospheric Dispersion Models

Abstract. Mapping of air quality are often based on ground measurements using gravimetric and air portable sensors, remote sensing methods and atmospheric dispersion models. In this study, Geographic Information Systems (GIS) and geostatistical techniques are employed to evaluate coarse particulate matter (PM10) concentrations observed in the Central Business District of Baguio City, Philippines. Baguio City has been reported as one of the most polluted cities in the country and several studies have already been conducted in monitoring its air quality. The datasets utilized in this study are based on hourly simulations from a Gaussian-based atmospheric dispersion model that considers the impacts of vehicular emissions. Dispersion modeling results, i.e., PM10 concentrations at 20-meter interval, show that high values range from 135 to 422 μg/mm3. The pollutant concentrations are evident within 40 meters from the roads. Spatial variations and PM10 estimates at unsampled locations are determined using Ordinary Kriging. Geostatistical modeling estimates are evaluated based on recommended values for mean error (ME), root mean square error (RMSE) and standardized errors. Optimal predictors for pollutant concentrations at 5-meter interval include 2 to 5 search neighbors and variable smoothing factor for night-time datasets while 2 to 10 search neighbors and smoothing factors 0.3 to 0.5 were used for daytime datasets. Results from several interpolation tests indicate small ME (0.0003 to 0.0008 μg/m3) and average standardized errors (4.24 to 8.67 μg/m3). RMSE ranged from 2.95 to 5.43 μg/m3, which are approximately 2 to 3% of the maximum pollutant concentrations in the area. The methodology presented in this paper may be integrated with atmospheric dispersion models in refining estimates of pollutant concentrations, in generating surface representations, and in understanding the spatial variations of the outputs from the model simulations.

scholarly journals Ocean Plastic Assimilator v0.2: assimilation of plastic concentration data into Lagrangian dispersion models

2021 ◽  
Vol 14 (7) ◽  
pp. 4769-4780
Author(s):  
Axel Peytavin ◽  
Bruno Sainte-Rose ◽  
Gael Forget ◽  
Jean-Michel Campin
Keyword(s):  
Dispersion Model ◽  
Simulated Data ◽  
Circulation Model ◽  
Particle Dispersion ◽  
Dispersion Models ◽  
Concentration Data ◽  
Suitable Candidate ◽  
Real World Applications ◽  
Lagrangian Dispersion ◽  
Assimilation Scheme

Abstract. A numerical scheme to perform data assimilation of concentration measurements in Lagrangian models is presented, along with its first implementation called Ocean Plastic Assimilator, which aims to improve predictions of the distributions of plastics over the oceans. This scheme uses an ensemble method over a set of particle dispersion simulations. At each step, concentration observations are assimilated across the ensemble members by switching back and forth between Eulerian and Lagrangian representations. We design two experiments to assess the scheme efficacy and efficiency when assimilating simulated data in a simple double-gyre model. Analysis convergence is observed with higher accuracy when lowering observation variance or using a circulation model closer to the real circulation. Results show that the distribution of the mass of plastics in an area can effectively be improved with this simple assimilation scheme. Direct application to a real ocean dispersion model of the Great Pacific Garbage Patch is presented with simulated observations, which gives similarly encouraging results. Thus, this method is considered a suitable candidate for creating a tool to assimilate plastic concentration observations in real-world applications to estimate and forecast plastic distributions in the oceans. Finally, several improvements that could further enhance the method efficiency are identified.

scholarly journals Inter-comparison of Atmospheric Trace Gas Dispersion Models: Barnett Shale Case Study

2018 ◽  
Author(s):  
Anna Karion ◽  
Thomas Lauvaux ◽  
Israel Lopez Coto ◽  
Colm Sweeney ◽  
Kimberly Mueller ◽  
...  
Keyword(s):  
Dispersion Model ◽  
Methane Emissions ◽  
Gas Production ◽  
Estimation Methods ◽  
Trace Gas ◽  
Barnett Shale ◽  
Dispersion Models ◽  
Data Set ◽  
Gas Dispersion ◽  
Tracer Dispersion

Abstract. Greenhouse gas emissions mitigation requires understanding dominant processes controlling fluxes of these trace gases at increasingly finer spatial and temporal scales. Trace gas fluxes can be estimated using a variety of approaches that translate observed atmospheric species mole fractions into fluxes or emission rates, often identifying the spatial and temporal characteristics of the emissions sources as well. Meteorological models are commonly combined with tracer dispersion models to estimate fluxes using an inverse approach that optimizes emissions to best fit the trace gas mole fraction observations. One way to evaluate the accuracy of atmospheric flux estimation methods is to compare results from independent methods, including approaches in which different meteorological and tracer dispersion models are used. In this work, we use a rich data set of atmospheric methane observations collected during an intensive airborne campaign to compare different methane emissions estimates from the Barnett Shale oil and natural gas production basin in Texas, U.S.A. We estimate emissions based on a variety of different meteorological and dispersion models. Previous estimates of methane emissions from this region relied on a simple model (a mass balance analysis) as well as on ground-based measurements and statistical data analysis (an inventory). We find that in addition to meteorological model choice, the choice of tracer dispersion model also has a significant impact on the predicted downwind methane concentrations given the same emissions field. The dispersion models tested often under-predicted the observed methane enhancements with significant variability between different models and between different days. We examine possible causes for this result and find that the models differ in their simulation of vertical dispersion, indicating that additional work is needed to evaluate and improve vertical mixing in the tracer dispersion models commonly used in regional trace gas flux inversions.

scholarly journals Dense gas dispersion model development and testing for the Jack Rabbit II phase 1 chlorine release experiments

2018 ◽  
Vol 192 ◽  
pp. 218-240 ◽  
Author(s):  
Simon Gant ◽  
Jeffrey Weil ◽  
Luca Delle Monache ◽  
Bryan McKenna ◽  
Maria M. Garcia ◽  
...  
Keyword(s):  
Dispersion Model ◽  
Phase 1 ◽  
Model Development ◽  
Dense Gas ◽  
Gas Dispersion

Imitation of Demonstrations Using Bayesian Filtering With Nonparametric Data-Driven Models

2017 ◽  
Vol 140 (3) ◽  
Author(s):  
Nurali Virani ◽  
Devesh K. Jha ◽  
Zhenyuan Yuan ◽  
Ishana Shekhawat ◽  
Asok Ray
Keyword(s):  
Particle Filtering ◽  
Dynamic Models ◽  
Training Data ◽  
Data Driven ◽  
Conditional Density ◽  
Bayesian Filtering ◽  
State Dependent ◽  
State Evolution ◽  
And Control ◽  
Nonparametric Kernel

This paper addresses the problem of learning dynamic models of hybrid systems from demonstrations and then the problem of imitation of those demonstrations by using Bayesian filtering. A linear programming-based approach is used to develop nonparametric kernel-based conditional density estimation technique to infer accurate and concise dynamic models of system evolution from data. The training data for these models have been acquired from demonstrations by teleoperation. The trained data-driven models for mode-dependent state evolution and state-dependent mode evolution are then used online for imitation of demonstrated tasks via particle filtering. The results of simulation and experimental validation with a hexapod robot are reported to establish generalization of the proposed learning and control algorithms.

Sign in / Sign up

Export Citation Format

Share Document