Peter Marx, RJM Corporation, USA, Robert Kish, Penn Power New Castle Generating Station, USA

Penn Power’s New Castle generating station has been brought in line with Clean Air Act requirements through a combination of CFD modelling and the upgrade of the boiler’s burners with customized parts. The result is lower NOx and improved efficiency – all for around one-third the cost of total burner replacement.

A reduction in NOx of more than 40 per cent, improved combustion efficiency, along with savings of 66 per cent over a costly burner replacement project. These objectives were achieved on Unit 5 at Penn Power’s New Castle generating station in New Castle, Pa., USA.

This 439 992 kg/h (970 000 lbs/hr) coal fired boiler was brought into compliance with Clean Air Act Amendment requirements by employing RJM Corporation’s computational fluid dynamic (CFD) modelling to locate combustion and opacity inefficiencies in the unit. The fix: RJM customized parts for the boiler’s 16 burners and followed up with optimization testing. For almost one-third of the cost of a total burner replacement or installation of new low NOx burners, Penn Power achieved dramatic results from the upgrade – a 40 per cent reduction in NOx, improved combustion efficiency, and a decrease in fly ash loss on ignition (LOI).

The Penn Power unit is a Babcock & Wilcox boiler that began operating in 1964. The unit has 16 burners arranged in four rows of four burners each. The furnace is balanced draft with induced draft and forced draft fans.

Plant engineers had originally conducted some burner tuning to lower baseline NOx from 0.5 kg/mmBtu (1.1 lb/mmBtu) to 0.36 kg/mmBtu (0.785 lb/mmBtu). Baseline fly ash LOI ranged from 12 per cent to 18 per cent. The utility’s objective was to reduce the unit’s NOx to below 0.2 kg/mmBtu (0.45 lb/mmBtu) throughout the load range with 12 per cent LOI.

The testing process

RJM developed and implemented a multi-step recommendation that included burner components, initial optimization, final optimization and coal flow testing. All the components and new parts installed on Unit 5 were custom fabricated by RJM subcontractors to function within the existing burners.

Burner installation: The goal of this portion of RJM’s work was to design a burner that could provide efficient combustion during most of the year and low NOx operation through the ozone period – the summer months of June through September. RJM used sophisticated CFD modelling on one of the burners and associated flow path to determine how the RJM components would affect NOx and fly ash LOI level. Modelling predicted the burner components would produce a 45 per cent reduction in NOx and maintain fly ash LOI at current levels.

During a two-week period in June of 1998, RJM retrofitted all 16 burners by adding a flame stabilizer, coal spreader and coal flow distributor. The stabilizer was designed to divide the secondary air into two zones to radially stage combustion. The coal spreader split the coal into three streams to create fuel rich and lean zones at the burner outlet. To create a uniform primary air and fuel mixture at the burner inlet, RJM installed a coal distributor upstream of the coal spreader.

Initial optimization: Once the unit was re-started, the objective was to optimize the burners to provide low NOx performance and limit opacity. Adjusting air registers, as well as coal spreader position and orientation optimized the burners. Graphs were developed to illustrate the relative effect of a single parameter on NOx formation while holding the other parameters constant.

Click here to enlarge image

Figure 1 shows NOx versus coal spreader position. Air registers were all set at 30 per cent open, excess O2 at 3.3 per cent, load at 130 MW and spreader oriented with a fuel rich zone 30 degrees from top dead centre. No fly ash LOI testing was performed during this phase.

Figure 2 shows results of the spreader rotation. RJM designed the flame stabilizers to create rich air and lean secondary air zones, with rotation that forces the coal into these zones. NOx changed from 0.25 to 0.23 kg/mmBtu (0.54 to 0.51 lb/mmBtu) when moving coal from the air rich to the air lean zones.

Click here to enlarge image

Final optimization and coal flow testing: The unit’s four pulverizers supply coal to the boiler elevations in a bottom-to-top arrangement of A, D, B and C. Tests determined that any combination of pulverizer operation that included the B pulverizer resulted in opacity spiking at part load.

RJM found that when the C pulverizer was removed from service, the effect of the coal flow imbalance of the next two pulverizers became evident. In addition, RJM found that when the unit was ramped down through the load range, and the upper pulverizer C was removed from service, opacity spiking became worse.

Various pulverizer combinations were tried and the B pulverizer was found to have the greatest impact on opacity.

Before RJM could determine final burner settings, Penn Power wanted to have the coal flows in all pulverizers balanced. Therefore, Penn Power plant engineers spent several days optimizing the pulverizer and burner settings before RJM completed a final burner optimization.

Penn Power continued to test pulverizers B and D several times and the coal flow imbalances exceeded ±25 per cent on B and D elevations. Penn Power hoped to achieve a balance within ±ten per cent.

RJM also performed part load tests and experimented by having various pulverizers in or out of service. When the B pulverizer was left in service and any of the other three pulverizers were removed, unit opacity would spike. It was determined that back end excess O2 and CO were severely out of balance, causing LOI to increase. Removing the B pulverizer and operating with the other three pulverizers resulted in no opacity spikes and balanced back-end O2 and CO. Other tests were performed on the B and D pulverizers.

Final fix

During the final optimization and coal flow testing phase, the unit was tested at 140 MW and excess O2 at three per cent. Initial readings were taken with the register at 25 degrees. NOx was at 0.19 kg/mmBtu (0.42 lb/mmBtu) and 9.5 per cent LOI. Superheater outlet steam temperatures could be controlled through an acceptable range with boiler soot blowing. RJM then opened the registers to 100 per cent, which increased NOx to 0.27 kg/mmBtu (0.6 lb/mmBtu), while LOI decreased to 5.9 per cent. The air registers controlled the mixing dynamics of the burner. Staging of the burner could be varied to provide optimum combustion or low NOx operation.

The final settings for the burners were as follows:

  • Register settings: 25 per cent
  • Spreader position: 15.2 cm forward on top two rows of burners and 0 cm forward on the bottom two rows.
  • Spreader orientation: at the two o’clock orientation

As a result of RJM’s burner components, NOx was reduced by 40 per cent and LOI by 24 per cent compared to the pre-construction baseline of 0.36 kg/mmBtu (0.79 lb/mmBtu) NOx with 16 per cent LOI. Table 1 provides a summary of the other conclusions of RJM’s burner upgrade and optimization programme.

Since the successful installation at Unit 5 in 1998, three other Penn Power boilers have received RJM burner components: two coal-fired boilers at the Sammis Generating Station of Ohio Edison in Stratton, Ohio, and one additional boiler at the New Castle Generating Station. In each instance, programme goals were achieved with similar results.

RJM burner programme at Unit 5, New Castle Generating Station

  • NOx ranged from 0.2-0.27 kg/mmBtu (0.44-0.60 lb/mmBtu) when changing the air register setting from 25 per cent to 100 per cent open. Fly ash LOI also changed from 9.5 per cent to 5.9 per cent over the air register range.
  • Opacity spikes occurred when the B pulverizer was brought into service. Subsequently opacity was brought under control with burner tuning and pulverizer biasing.
  • NOx emissions rate changed by 0.032 kg/mmBtu (0.07 lb/mmBtu) depending on the position of the coal spreader.
  • NOx emissions ranged from 0.21-0.18 kg/mmBtu (0.47-0.40 lb/mmBtu) when excess O2 was changed from 3.3 per cent to 2.2 per cent.
  • Burner 5B and the D pulverizer each experienced coal flow distributions exceeding ±25 per cent.
  • Superheater outlet temperatures were controlled by soot blowing cycles
  • NOx was reduced by six per cent by rotating the coal spreader 60 degrees.