Pollution Prediction Model Combining CFD and Gaussian Simulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for predicting pollution diffusion in complex terrains with obstacles like industrial parks are either inaccurate due to ignoring obstacles or too slow due to computational fluid dynamics simulations, making timely predictions challenging.
Innovation Solution
A method and apparatus using a target available model trained with a combination of fluid dynamics and Gaussian simulation models to generate training samples from predetermined environment data, allowing for fast and accurate prediction of pollution concentration sequences and evacuation routes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If computational fluid dynamics simulation is used to predict pollution diffusion in complex terrain, then prediction accuracy is improved, but computation speed becomes very slow
Solution Approach 1:
The patent pre-trains a machine learning model using computational fluid dynamics simulation data before actual prediction needs arise. This preliminary training allows the model to capture complex terrain pollution diffusion patterns in advance, so that during actual use, the model can make fast predictions without running time-consuming CFD simulations again.
Solution Approach 2:
The patent creates a simplified machine learning model that copies and learns from the complex computational fluid dynamics simulation data. Instead of directly running the complex CFD simulations for every prediction, the ML model creates a faster copy that replicates the essential pollution diffusion behavior, achieving both accuracy and speed.
2Productivity
If Gaussian simulation model is used for pollution diffusion prediction, then computation speed is fast, but prediction accuracy deteriorates due to ignoring obstacles and buildings
Solution Approach 1:
The patent introduces a machine learning model as an intermediary that bridges the gap between simple Gaussian simulations and complex CFD simulations. The ML model learns from detailed CFD data and incorporates the effects of obstacles and buildings, serving as a mediator that provides both speed and accuracy.
Solution Approach 2:
The patent transforms the prediction approach by changing from using physical parameters directly (as in Gaussian or CFD models) to using learned parameters from training data. The machine learning model processes input parameters (source strength, wind speed, direction) and outputs prediction results that accurately reflect complex terrain effects without requiring explicit physical modeling of obstacles.
Data Source
AI summary
Various embodiments of the teachings herein include an environment prediction method based on a target available model. An example method comprises: generating a training sample based on predetermined environment data; using the training sample to perform training based on a fluid dynamics model and a Gaussian simulation model, to obtain a target available model; and based on real environment data, using the target available model to determine a real environment prediction value of a time-related pollution concentration sequence for a calibration position.