The BP refinery in Whiting, IN has 80,000 valves in its leak detection and repair (LDAR) program. There are approximately 20 process units in the refinery, on an average five-year-turnaround cycle and 200-300 valves on every turnaround work list. BP executed an EPA consent decree in 2012 and has been working on improvements around LDAR and emissions since 2011.
Analyzing the Problem
There are several failure modes for valves, including bonnet and flange leaks, and leaks through the seat, but 77% of them are due to stem packing leaks. BP conducted a number of tests during the 1990s to obtain an understanding of the problem and explore possible solutions.
Tests on rising stem gate and globe valve packings showed that all graphite packings perform better than asbestos. Some relatively simple low- to medium-density graphite packings with braided end rings can give very good performance, and graphite packings generally are relatively insensitive to stem damage. However, the best performance was obtained from an "engineered" design of packing with some pressure-energizing capability.
Stem straightness and run-out were also analyzed with the result that, when evaluating a valve manufacturer, their manufacturing practices were analyzed to see how well they could control straightness. The gland follower must be guided on the packing chamber, not the stem.
Also, while good surface finish and close tolerances may give improved sealing performance, very smooth stem finishes produced by burnishing are actually detrimental to good sealing of graphite packings.
Before the Consent Decree
Prior to the consent decree, BP Whiting was also focusing on chronic leakers or those that created problems for unit personnel. Efforts to solve the problems included maintenance that involved tightening gland nuts, repacking, reconditioning and pumping up. When even that did not solve many of the problems, energies focused on being more rigorous with re-tightening, using more reconditioned valves and repacking with live-loading.
Other strategies included replacement of the valve, but, since 77% of the failures are stem packing leaks, it seemed futile to buy a whole new valve, even if it was from screened manufacturers using new technologies. It’s like buying a new car when you have a flat tire!
It’s essential to consider the life-cycle cost of the installed valve. Before the consent decree, it was common for BP to replace valves, but replacing them was not producing better results in reducing emissions. In one light-end unit about 20% of the valves replaced in 2007 were on the work list again in 2012, and many had leaks within the first year.
BP Whiting tried to improve the performance of new valves by requesting testing from its suppliers. The level of testing and cooperation was mixed, ranging from nothing to very thorough. Even when testing was performed, there were numerous variables to decipher such as:
- Packing or valve test
- In house, supervised/certified or outside lab testing
- Valve randomly selected
- Test type (API 622, ISO 15848, company tests, etc.)
- Emissions limit (100 ppm or 500 ppm)
- Test attributes (cycles, temperature, pressure, media, etc.)
Meeting the Consent Decree
Once the consent decree was in place, there were new specifications for purchasing valves. Valves now must be certified low-leaking valves or repacked with certified low-leaking packing. These are defined as “valves for which a manufacturer has issued either a written guarantee that the valve will not leak above 100 parts per million (ppm) for five years or a written guarantee, certification or equivalent documentation that the valve has been tested pursuant to generally-accepted good engineering practices and has been found to be leaking at no greater than 100 ppm.”
The guarantee has to be via a letter, not just a claim on a website or in a brochure. To ensure that valves met all of the consent decree requirements, Whiting created its own specification: WBU-GIS 62-1001 states that rising stem valves require successful testing, but that quarter-turn valves require only a guarantee.
To create a list of approved valves, BP Whiting representatives went to more than 200 valve manufacturers that were in the company’s procurement records and asked for documentation and testing. It was essential to get this information quickly, as we needed to buy valves, so a methodology was developed to analyze and judge performance of various test data. It was then added to BP’s specification when it was finalized.
The attributes used to judge the different tests from suppliers included both comparing to standards and comparing valves to each other and other companies. Ratings considered include:
- Performing entity (an independent lab)
- Mechanical cycles
- Thermal cycles
- Maximum temperature
- Maximum leakage rate (ppm)
- Mean leakage rate (ppm)
- Gland nut adjustments performed/allowed
The results were scored and a list of approved vendors with ratings was created. When purchasing new valves, this is one of the facets of the plan to achieve our goal of minimizing fugitive emissions.
In addition to being proactive in approving manufacturers, BP Whiting has embraced new packing technology and new valve technology. For example, bellows sealed valves might be considered for services that have a documented history of leakage or for thermal/mechanical cycling services.
When repacking and repairing valves, only internally approved materials are used. When valves are repacked there is a detailed examination of stuffing box components including clearances, surface finishes and packing dimension; and during new packing installation the sequencing and torque is inspected.
While it is relatively early to judge progress, early results have shown a 21% decrease in one unit’s valve leaks. There have been no bellows leaks. Previously, 20% of new valves leaked in the first year, but it has been reduced by 180 less leaks in one year in just one unit. The overall leak rate has decreased from over 30% between 2008 and 2013.
Overall, BP Whiting’s performance has been trending positively, even though the new measures have been followed for only a couple of years.
The Path Forward
We will continue to track performance using the LDAR database and plans have been made to improve the database to make it friendlier for reliability work. We will also judge performance of more reliable solutions (life-cycle analysis), whereby we tighten, re-monitor, pump-up, procure, remove and install prematurely on some percentage of valves as compared to replacing with a longer life (higher MTBF) alternative.
We also plan to request valve suppliers submit API 624 test results for our review and will continue to focus on chronic leakers within units, enhance the reconditioning program and continue to make improvements in our quality assurance program for valves.