If you’re in the coal-fired power generation business, you may know that Coal-Gen 2009 is going on this week. During the conference, a Midwest power producer and the Emerson Power & Water Solutions team gave a joint presentation on combustion optimization.
Emerson’s Jeff Williams, one of the presenters, was kind enough to send me a copy so I could relay a few highlights in this post. The presenters discussed how they were able to optimize the combustion process to reduce NOX levels beyond the guarantee level.
Coal-fired power plants are impacted by many dynamic factors including source fuel type & quality, market deregulation, tightened emission standards to name a few. Costs for NOX and SO2 credits have increased over the last few years.
There are many pre- and post-combustion technologies available to reduce NOX and SO2 emissions, each with its own cost-benefit ratio–investment cost of the technology vs. the %NOX reduction.
For the project described in the presentation, the team benchmarked pre-project NOX, O2, and steam temperature levels and burner tilt performance. Two improvements were identified, the addition of separated OverFire air (SOFA) dampers & tilts and combustion optimization in the plant’s Ovation control system.
The OverFire air process redistributes air within the boiler combustion zone and injects additional air above the combustion zone to complete the combustion process. Decreasing the air within the burner zone lowers stoichiometry, which lowers the flame temperature and reduces thermal NOX. This also reduces the tendency of fuel-bound nitrogen to oxidize to nitrous oxides.
To compensate for temperature excursions caused by rapid changes in SOFA positions, advanced control strategies were developed. These control strategies were based on an advanced non-linear, fuzzy-neural NARMAX (FNM) algorithm.
The team followed a multi-step process, which included a study of the current combustion process, DCS control improvements, parametric testing, model development, open-loop testing, closed-loop testing, and commissioning.
For this project’s optimization model NOX and CO were the control variables. Manipulated variables included the OFA and SOFA dampers, SOFA tilts, O2 trim, auxiliary air dampers, window-to-furnace differential pressure, fuel air dampers, and feeders. The disturbance variables included load, ambient temperature, total air flow, and burner tilts demand.
Over the multi-year process that included the combustion optimization, followed by the SOFA equipment, followed by the advanced control optimization of the SOFA equipment, the plant reduced annual NOX output from over 1400 tons to under 600 tons.