ENV-17-28 Refinery Process Stream Composition Data – Update to Speciation Data in API 4723

Description:

The Petroleum Environmental Research Forum (PERF) previously completed a project in the 1990’s to create a compilation of refinery process stream speciation data, which was later published as API 4723, Refinery Stream Speciation (API 4723). The refining industry has routinely used these data to estimate hazardous air pollutant (HAP) fugitive emissions, storage tank and miscellaneous process vent emissions, and to prepare Toxic Release Inventory reports. Since API 4723 was published, regulatory and other requirements (e.g. Mobile Source Air Toxics standards, reformulated gasoline standards, lower fuel sulfur standards) have resulted in refinery process changes that potentially altered compound speciations in process streams. Changing crude feedstocks, new process additives, and new catalysts have also affected what species are present in specific process streams. PERF has initiated a new project to re-evaluate current process stream speciation data from several refineries. Seventy-one refinery process streams were characterized using an expanded list of chemical species or analytes. These analytes are potentially found in refinery process streams and identified in Hazardous Air Pollutant, Persistent Bioaccumulative Toxic, or Toxic Release Inventory lists. Up to 73 analytes were included per process stream compared to the 24 analytes in the original API 4723 study. Most of the new analytes are metal HAPs and heavier semi-volatile compounds. Preliminary project results indicate decreases in chemical species “Most Likely Values” (MLV) for many process streams versus the original study values. Some process streams had increases but these results were generally based on larger sample data sets supporting them. More compounds may have been identified in a given process stream compared to the previous study due to the expanded list of analytes or to improved laboratory methods providing lower method detection limits (MDLs).