The Importance of Tracking Hydrogen H2 in Complex Natural Gas Networks

Abstract

Abstract Objectives/Scope Many midstream operators are developing plans to introduce Hydrogen (H2) into natural gas networks in concentrations of 0 to 20%. It is important to keep the concentration in all locations in the network within these limits, to reduce H2 induced cracking of pipelines and because most burners in the US are not designed for the different heating values (Wobbe Index) associated with the H2 mixtures, and flame speed present with H2 flames. This paper presents a model which dynamically tracks H2 concentrations, and associated Wobbe Index, in a complex delivery network and discusses the operational challenges associated with the introduction of H2 in these systems. Methods, Procedures, Process The model is based on proven technology which has been used on gases of variable quality for 30 years. The key item currently however is the new addition of H2 to the composition mix. The model is dynamic and responds to changes in the H2 concentration and flowrate. It is envisioned that the H2 supply to these networks will be variable (for example solar generated H2). This paper will present several case studies showing how H2 affects an existing gas distribution network and the paper will also discuss how H2 impacts the thermodynamics used in the calculation engine. Results, Observations, Conclusions The case studies will show that for even simple events like a customer trip, high concentration H2 packets can travel into pipelines that normally do not receive H2. As H2 travels through the pipeline, parameters within the system start to change such as the pipeline pressure drop increases and the compressor duty and outlet temperature increase. Novel / Additive Information The model tracks gas packets dynamically in the network (packet size is based on dispersion length) and calculates the density and energy content based on the local concentration. It then has a built in EOS (equation of state) based on GERG 2008 to calculate the density. Even with this complexity the model can run at speeds 100 times real-timefor a network that has approximately over 1,000 km of pipe.