The sulphur deadline

Paul Boughton

Compliance with new EPA regulations for refinery flares will soon be mandatory. Ben Hill asks whether your plant is ready

The three-year grace period many refineries have enjoyed is coming to an end. When promulgated in 2012, the final EPA New Source Performance Standards (NSPS) subpart Ja only applied to new flares built after June 24, 2008. But now, in only a few short months, NSPS Ja will be expanded to cover all flares modified after June 24, 2008. Because the list of flare piping changes that don’t count as modifications is so short, this effectively means that all US refinery flares must comply by November 2015.

The expanded regulation will have a major impact on flare operation. In particular, refineries subject to the regulation will now be required to continuously monitor total reduced sulphur (TRS) and provide detailed and accurate reports on sulphur emissions to regulators. For many, this will mean installing a sulphur online analyser that can measure total sulphur content in real time. When choosing an analyser, plant managers should consider several key factors and capabilities – most notably, the analyser’s measurement technique, dynamic range and recovery time.

Two primary techniques exist for measuring total sulphur in flare emissions: pulsed ultraviolet fluorescence (PUVF) and gas chromatography (GC). GC is a well-known technique in laboratory settings, and is fairly common for refinery sulphur monitoring. GC instruments determine the composition of a gas by passing it through a column filled with an inert carrier gas (often helium). As the sample passes through the column, each component is slowed down to a different degree by the carrier gas. By measuring how long it takes each component to reach a detector at the end of the column, the instrument can determine the sample composition.

Although GC is excellent for measuring the contents of a single, well-defined gas phase sample, it is less effective with continuous measurement. A process GC has an analysis cycle that can take several minutes to complete, which can cause a blind period where a flare event could occur. Furthermore, as samples are fed continuously into the sample injector, GC instruments can accidentally confuse gases slowed during a previous sample with gases from the current sample. GC column carrier gas also slows different sulphur containing molecules to different degrees, which is another potential source of total sulphur measurement inaccuracies. Problems can also arise as the column degrades over time.

Unlike GC analysers, instruments that use PUVF measure sulphur by irradiating flare emissions with pulses of a specific wavelength of ultraviolet light. This light excites the sulphur, and the intensity of the emitted light is directly proportional to the sulphur concentration of the flare stack emission and can therefore be used to calculate total sulphur. Because PUVF uses light rather than carrier gas to measure the presence of sulphur and uses fewer consumables than GC, its operating cost is typically lower. This technique is especially effective when emissions are combusted into pure SO2 before being analysed, as this eliminates the necessity to monitor more than one sulphur-containing molecule.

Dynamic range

The majority of flare stacks have sulphur levels that vary greatly from flare event to flare event. This means that flare gas analysers must be able to measure a broad, dynamic range of sulphur. Although this has traditionally been achieved by simply broadening the detection limits of a single measuring range, newer approaches favour using two independent measuring ranges, creating a more robust and less expensive solution.

By using dual sample injection valves with a 100:1 dilution ratio and dual photomultiplier tubes (PMTs) to set two different detector sensitivity levels, analysers with dual measurement ranges can measure the same sample gas using either a 0(x) to 1(x) range (for example, 0-3,000 ppmv) or a 1(x) to 100(x) range (3,000-300,000 ppmv). In this example, the two detectors would deliver a total dynamic range of 0-300,000 ppmv. A high-quality system should, however, be able to measure from 10 ppmv to 100% sulphur by volume before being configured.

Selection of the measurement range and sample injector is done by software, automatically and in real-time. This means that the actual amount of sulphur reaching the detector is relatively constant – during high sulphur periods, the analyser will simply switch to the higher measurement range. Because the instrument’s detector no longer has to handle large variations in sulphur levels, measurement repeatability and instrument uptime are significantly increased. This has dramatic effects on another key aspect of successful NSPS subpart Ja compliance: instrument recovery time.

Recovery time

To reliably monitor a flare stack and meet the requirements of NSPS subpart Ja, a sulphur online analyser must have a recovery time that is shorter than the minimum period between the stack’s flare events. Most monitoring instruments struggle with this because their detectors take a long time to recover from measurements in percentage levels to ppm levels. As previously shown, dual-measurement solutions solve this problem by combusting and diluting high sulphur samples before they reach the detector, providing a much less variable total sulphur content. Sulphur online analysers using this technology can easily recover from a high sulphur sample in as few as eight minutes. Some operators have even reported recovery times in the five-minute range.

In addition to flare events, reduced recovery time has benefits for instrument validation. Once every 24-hour period, all NSPS-compliant refineries must use a high sulphur standard that is within 80-100% of the analyser’s calibrated high measuring range to validate the performance of their instrument. This requires instrument downtime – and because regulation dictates that each validation gas be analysed for 10 minutes, one of the only options plant managers have for reducing this downtime is minimising instrument recovery time. Because this validation process happens daily, small efficiency gains and saved minutes quickly become saved hours over the course of a quarter – and, over the course of a year, saved days.

Regulations are rarely welcomed – they can disrupt business, generate costs and complicate operations. But disruption, cost and complication aren’t absolute. As NSPS subpart Ja expands to cover all flare stacks, plant managers do indeed have proven options for reducing the impact on their businesses. But first they must carefully consider the necessary infrastructure required for the new continuous TRS monitoring. And starting with a thorough assessment of measurement technique, dynamic range and recovery times of their sulphur online analysers is a wise decision.

Ben Hill is applications manager, environmental and process monitoring, Thermo Fisher Scientific.

Recent Issues