Low-Cost Representative Sampling for a Natural Gas Distribution System in Transition

Abstract

Natural gas distribution systems within municipalities supply a substantial fraction of energy consumed in the United States. As decarbonization of the natural gas system necessitates new modes of operation outside original design purposes, for example, increased hydrogen or biogas blending, it becomes increasingly important to understand in advance how existing infrastructure will respond to these changes. Such an analysis will require detailed information about the existing asset base, such as local soil composition, plastic type, and other characteristics that are not systematically tracked at present or have substantial missing data. Opportunistic sampling, for example, collecting measurements at assets that are already undergoing maintenance, has the potential to substantially reduce the cost of gathering such data but only if the results are representative of the full asset base. To assess prospects for such an approach, we employ a dataset including the entire service line and leak database from a large natural gas distribution utility (∼66,700 km of service pipelines and over 530,000 leaks over decades of observations). This dataset shows that service lines affected by excavation damage produce an approximately random sample of plastic and steel service lines, with similar distributions of component age, operating pressure, and pipeline diameter, as well as a relatively uniform spatial distribution. This means that opportunistic measurements at these locations will produce a first-order estimate of the relative prevalence of key characteristics across the utility's full asset base of service lines. We employ this approach to estimate the plastic type, which is unknown for roughly 80% of plastic service lines in the database. We also find that while 32% of leaks across all components occur in threaded steel junctions, excavation damage accounts for 75% of hazardous grade 1 leaks in plastic service lines and corrosion accounts for 47% in steel service lines. Insights from this sampling approach can thus help natural gas utilities collect the data they need to ensure a safe and reliable transition to a lower-emission system.