Updates to Requirements for Heat Exchange Systems at Petroleum Refineries

On June 20, 2013, the U.S. EPA finalized several changes regarding heat exchange system requirements in 40 CFR 63 Subpart CC – National Emission Standards for Hazardous Air Pollutants from Petroleum Refineries (MACT CC).  The rule was reevaluated in response to a petition filed by the American Petroleum Institute (API), although the revisions go beyond the issues raised by API.

The heat exchange system new source definition cut-off was changed from August 18, 1995 to July 14, 1994, which matches the general MACT CC new source definition.  Other major changes include updates to the definition of “heat exchange system,” the addition of a reduced monitoring frequency option for existing sources, and additional monitoring options and clarifications.  Note that the proposed rule had created a national “Uniform Standard” for heat exchange system, with MACT CC referencing the standard.  EPA decided not to pursue the Uniform Standard at this time and chose to directly amend MACT CC instead. 

The heat exchange system definition was changed by replacing the phrase “series of devices” (used to transfer heat from process fluids to water without direct contact and to transport and/or cool the water in the cooling system) with “collection of devices.”  Additionally, sample coolers and pump seal coolers have been explicitly excluded from the definition (EPA indicated that they had never intended to regulate these devices).  Lastly, EPA clarified that only individual heat exchangers “in organic HAP service” and servicing a “petroleum refinery process unit” are subject to the rule.  (Previously, MACT CC indicated that entire heat exchange systems, not individual heat exchangers, must meet these applicability criteria.) 

Additionally, throughout the rule, references to “VOC” or simply “emissions” have generally been replaced by “total strippable hydrocarbons.”  This phrase more closely matches the data provided by the Modified El Paso Method.  Note also that EPA chose not to finalize the option to use direct water sampling methods, which had been included in the proposed Uniform Standard, as an alternative to the Modified El Paso Method. 

In general, MACT CC requires monthly El Paso sampling with a leak detection level of 6.2 ppmv as methane.  The revisions now allow existing sources to sample quarterly instead with a leak detection level of 3.1 ppmv, with a switch to monthly monitoring if a leak repair is delayed.  One monitoring option must be chosen per heat exchange system.  A refinery may switch between monitoring options, but first all leaks must be repaired and a notification must be submitted at least 30 days prior to the change.

For once-through cooling systems, flexibility in the monitoring location was added such that aggregation of multiple heat exchange systems is allowed.  Now once-through systems must be monitored at selected heat exchanger exit line(s) so that each heat exchanger or group of heat exchangers within a heat exchange system is covered by a monitoring location.  Monitoring at locations after multiple heat exchange systems are combined is allowed so long as the total cooling water flow rate at the location does not exceed 40,000 gallons per minute. 

Outlet monitoring for once-through cooling systems may be supplemented by monitoring at the inlet water feed line prior to any heat exchanger.  Using this inlet data, a leak may be defined by the difference between the inlet and outlet concentrations, rather than just the outlet concentration.  EPA added this option to acknowledge that a once-through system may be receiving contaminated water from the source pond, river, etc.  EPA did not add a differential leak definition option for closed-loop systems, noting that strippable hydrocarbons will largely be removed in the cooling tower.

EPA also clarified that closed-loop recirculation heat exchange systems may be monitored not only in the cooling tower return line or at the heat exchanger exit lines, but also in a representative riser within the cooling tower, so long as the sampling is conducted prior to exposure to air.

A statement was added indicating that leak “repair” includes re-monitoring to demonstrate that the concentration is now below the leak definition.  Additionally, the rule now indicates that if monitoring at a cooling tower return line indicates a leak, but subsequent monitoring at individual or groups of heat exchangers within the system does not indicate a leak, then the leak is considered to be “repaired.”

Some changes were also made to the requirements for calculating emissions during delay of repair.  EPA clarified that delay begins on the date the leak should have been repaired and that water flow rate can be determined based on various engineering methods (e.g., pump curves).  Additionally, calculations now must be made per monitoring interval and not simply for the entire duration of the delay period (to better align the calculations with the data received).

Lastly, several recordkeeping and reporting updates were made.  A five year recordkeeping duration was specified for heat exchange system records, and only heat exchangers in petroleum refinery process unit service (as opposed to all heat exchangers) must be identified.  Additionally, the date the source of a leak was identified must be recorded, and flow rate now must only be recorded if repair is delayed.  Records for delayed repairs now include flow rate estimates as well as estimated potential strippable hydrocarbon emissions per monitoring interval.  In the Periodic Report, the number of heat exchange systems subject to monitoring (rather than simply in HAP service) must be reported; “a summary of monitoring data” for leaks has been replaced with more specific requirements; post-leak repair re-monitoring results must be reported; and monitoring results for leaks for which repair has been delayed must be reported, along with the expected date of repair and an estimate of emissions for each monitoring interval during the delay of repair.