Experimental Measurement and Prediction of Methane Emission Factors: Case Study District Seven of Gas Transmission Operation

In gas transmission processes, a significant amount of methane gas is transferred to the environment for several reasons, including leakage and pipe breakage, which has a high contribution to greenhouse gas emissions. The global methodology for reducing greenhouse gases has led to a plan to reduce greenhouse gas emissions from gas transmission lines according to international conventions and the national action plan to reduce greenhouse gas emissions. Therefore, the purpose of this study in first was experimentally measurement of methane gas emission due to leakage of blowdown valves from different points of district seven of gas transmission operation and then the prediction of gas emission factors using multilayer perceptron neural network with backpropagation of error and Levenberg-Marquardt algorithm. Accordingly, the effect of operating parameters on the emission factor, including ambient temperature, pipeline pressure, type of valve, condition of the valve (open or close) located before of blowdown, Internal leakage of bypass valve before blowdown, life of blowdown valve, the sound of leakage and concentration of the leakage in the range of 10-10000 ppm were investigated. To find the effective parameters on the emission factor, six various schemes were considered based on different input parameters. The results of prediction of methane emission factors using neural network showed that the scheme with input parameters as ambient temperature, pipeline pressure, age of valve, the sound of leakage and concentration of the leakage shows the best results. Deviation with measurements and coefficient of determination (R squared) of the best scheme among the six schemes are equal to RMSE=0.05047, MSE=0.00255, and R2=0.97853, which they show that this method can be used to predict the methane emission factors in natural gas transmission pipelines.