论文部分内容阅读
城区中有毒气体突发性泄漏时,需要快速对泄漏源进行定位和识别,以便科学预测气体的蔓延及其影响范围。利用基于Bayes推断理论的MCMC(Markov chain Monte Carlo)抽样方法,根据城市中分布的传感器测量信息和气体扩散数值计算模型,构造似然函数,对泄漏源的位置、强度进行反演。计算了这些参数和空间各点浓度的相关统计量,表明反演结果与泄漏源的真实参数十分吻合。此外,还讨论了传感器测量误差的概率分布对结果的影响。结果表明,误差概率会显著影响计算效果,概率分布越平坦,泄漏源反演信息的不确定度越大。
Sudden leakage of poisonous gases in urban areas requires rapid location and identification of sources of leakage in order to provide a scientific forecast of the spread of the gas and its impact. Using Markov chain Monte Carlo (MCMC) sampling method based on Bayesian inference theory and according to the sensor measurement information and gas diffusion numerical calculation model distributed in the city, the likelihood function is constructed to retrieve the position and intensity of the leak source. The correlation statistics of these parameters and the concentration of each point in the space are calculated, which shows that the inversion results are in good agreement with the true parameters of the leakage source. In addition, the influence of the probability distribution of sensor measurement error on the result is discussed. The results show that the error probability will significantly affect the calculation results, the more flat the probability distribution, the greater the uncertainty of the leakage source inversion information.