Low NOx Retrofit on an Existing Heater with a Small Radiant Section

Paper from the AFRC 2014 conference titled Low NOx Retrofit on an Existing Heater with a Small Radiant Section by T. Gilmartin.

Conventional logic in retrofitting a heater to meet Low NOx environmental limits suggests that you should only fit a burner that just meets the NOx requirement. A main factor in this logic is flame size where the critical factor is length - low NOx burner flames get longer larger the lower the NOx target is. But is this always true?! Of course you can always fit any burner - but will you maintain the same fired duty? This paper describes: The history: Six years ago, BP asked for a deferment to retrofit low NOx burners, as they saw no existing burner type available that would not cause a rate reduction with a flame fitting in the fire box. The thought process: Emission limits require 100 mg/Nm3. Which burner, if any, could meet compliance requirement for retrofit activities? Burner picked: CUBL HC burner seen on test - potential to retrofit and not sacrifice rate? Duty of care: visit other user, CFD and test Steps taken as a result of CFD and testing: End burner modifications after CFD showed potential end wall issues. Commissioning: demonstrated performance at full rate. Test runs at load showed 33ppm NOx (66mg/NM3) • An accurate CFD model assisted in gaining confidence that the retrofit would not cause issues. Show CFD and video of test. Lessons: CFD gives confidence. Paradox - sometimes fitting an ultralow NOx burner will perform better than a low NOx burner.

1 Presentation Title Goes Here 2013 UOP Refining R&D Panel Meeting 1914 - 2014 A Century of Innovation in the Oil and Gas Industry © 2014 UOP LLC. All rights reserved. UOP 6123-1 Low NOx Retrofit On An Existing Heater With A Small Radiant Section Tom Gilmartin BP Kurt Kraus, PE Callidus Technologies, LLC UOP, A Honeywell Company  Successful retrofit of Ultralow NOx burners on an existing heater while maintaining the original design firing rate, outlet temperature and charge rate  New NOx requirement of 150mg/NM3  The Processes - Review existing heater size, data and field experience - Select the most appropriate burner - Perform CFD simulation to indicate flame dimension - Perform Burner Test to indicate NOx emissions - Validate CFD results via burner test - Install burners in field and perform field testing The Challenges 2 Presentation Title Goes Here 2013 UOP Refining R&D Panel Meeting Background The Equipment  Twin cell cabin vacuum heater with eight floor fired burners per cell with the burners operating in forced draft mode (100mm H20) and air preheat of 100C  Originally designed to fire oil and gas, the firing on the unit was restricted to gas only firing in 2000  Tubes laid out in a horizontal configuration with a double row of roof tubes above each cell. Above those roof tubes there is a layer of ceramic tiles designed to improved radiant heat transfer  A single convection section is mounted above the two cells in a center position  Burners rated at 4.5MW normal and 5.5MW design duty Heater Drawing View of 3D Rendering From CFD Model Based on 2D Drawings Radiant Cell Dimensions and Field Data  Height 6857mm  Width 3038mm  Known issues with flame wash on side and roof tubes with the existing burners Hot spots on tubes caused by flame wash on tubes with original burners Dimensions are limiting! 3 Presentation Title Goes Here 2013 UOP Refining R&D Panel Meeting Comparison Against Standards  API STD 560  Horizontal dimensions - Min clearance 1.59m horizontal - For a Low NOx Forced draft burner (5.5MW)  Internal standard - 50 percent more clearance required i.e. 2.35m Actual dimensions 1.515m Note: No guidance is given on height "They" - conscientious engineers project developers • Try to follow and comply with industry standards - requires Homework • What do the industry standards say? Are there gaps in standards? Excerpts From API STD 560 "Fired Heaters For General Refinery Services" Flame height 2/3 of the firebox Other Helpful Clauses To date, there is no guidance on flame height in API for forced draft burners, the linear relationship shown in API STD 560 does not apply • API assumes all low NOx burners act and perform the same, regardless of model or make " " " " 4 Presentation Title Goes Here 2013 UOP Refining R&D Panel Meeting To date, there is no guidance on flame height in API for forced draft burners The linear relationship shown in API STD 560 for natural draft burners just does not apply to forced draft burners Tables 13 and 14 below are excerpts from API STD 560: "Fired Heaters For General Refinery Services" The homework does not look promising... Doctrine says: "It can't be done." Doing One's Homework Burner Selection and Proving  The availability of a CUBL HC burner designed for high air pressure drop was identified early in the project - This burner was witnessed both at Callidus then at a location in Europe (though on a vertical cylindrical heater where flame diameter rather than height is at issue) - Good experience with the conventional natural draft variant of this burner  The above points were sufficient to start a project following our normal protocol of... - Carrying out a full CFD analysis - In tandem with an extensive burner test at the manufacturer 5 Presentation Title Goes Here 2013 UOP Refining R&D Panel Meeting Burner Testing  Burner testing with particular reference to flame dimensions and emissions - Since test furnaces are typically challenging to visually see the whole flame envelope (because of the size of the furnace), CO probe measurement was used to define these dimensions. - Note that 250-300ppm CO was used as the definition of edge of flame rather than the more conventional 2500ppm recommended in API535 - An operator can see flame haze (250ppm) on a real furnace. An operators job is to ensure that there is clear separation between tube and flame, hence the reduced value Since this is a ultralow NOx burner, little concern at this stage in meeting the NOx values Burner Test and CFD results  Criteria for NOx was to achieve 100mg/NM3 @3%O2 at 280C air temperature (future case). Achieved in burner testing was 59mg/NM3 at these test conditions  Flame limited to 4570mm 66% of box height at design heat liberation 6 Presentation Title Goes Here 2013 UOP Refining R&D Panel Meeting CFD Work Validated With Burner Testing  This CFD was validated with burner test results obtained so that an accurate model of the real furnace could be built with all the burners in place - Although a single burner is a good test, it is rarely accurate enough to give accurate flame dimensions in a multi-burner environment because of flame interaction, box configuration etc. - CFD also permits checking that the existing burner layout would be suitable - Unlike a new heater where layout and dimensions can be easily adjusted during design, altering the burner layout on an existing heater can be very costly, requiring major heater floor works Creating the 3D Model and Results This modeling indicated that at the end walls there was some unacceptable flame interaction. To counter this, turning vanes were introduced into the model on the end burners, which (at least in the CFD) greatly reduced this effect. 7 Presentation Title Goes Here 2013 UOP Refining R&D Panel Meeting Burners were installed in Nov. 2013, which was followed by commissioning and test runs. Visual flame shape was very close to that predicted by the CFD at design rate. NOx 60mg/NM3 achieved. Field Installation and Commissioning  Successful retrofitting of ultralow burners on an existing heater while maintaining the original design firing rate. New NOx requirement of 150mg/NM3  Historic perspective: "It can't be done..." - But it can…  If you have forced draft and correct methodology, you can reduce risk and deliver a more consistent flame to heat the process - Natural draft may be cheaper, but it is also less consistent, especially when applied to a "tight" firebox. It has a greater project and operational risk Summary Forced draft: More than 50mmH2O (2" W.C.) available across the burner. The CUBL-HC model burner is suitable for pressures greater than 200mmH2O (8" W.C.) Increased pressure only improves all performance indicators. 8 Presentation Title Goes Here 2013 UOP Refining R&D Panel Meeting UOP 6123-15