%0 Thesis %K Energy %K Air quality %K Energy System Planning & Grid Integration %K Electrical Dispatch %A Chris Marnay %B University of California at Berkeley %C Berkeley, CA %D 1993 %I University of California at Berkeley %T Intermittent Electrical Dispatch Penalties for Air Quality Improvement %8 01/1993 %9 Doctor of Philosophy %X

Like most large U.S. urban areas, the photochemical smog problem of the San Francisco Bay Area occurs as interm ittent episodes. Current electric utility emissions regulations rely on mandated stack clean-up equipment which performs poorly on interm ittent problems because emissions reductions are not concentrated during episodes. Electric utilities have some flexibility in the boilers they commit and dispatch during episodes, so a policy to encourage use of cleaner boilers could result in reduced emissions of the smog precursor gas NOx. Such a policy deviates from the tradition of utility minimum cost operations, resulting in higher fuel costs. In this study, the dispatch of more polluting thermal generation within the confines of the Bay Area Air Quality Management District (BAAQMD) is penalized by the imposition of a variable NOx tax. The effect is explored through a Lagrangian relaxation unit commitment and dispatch simulation of an historic episode in September 1989, using a Monte Carlo sampling of outage states. Imposition of the tax results in modest NOx reductions, achieved for a small increase in fuel bill. The cost per avoided ton of NOx emitted during the episode is low compared to the costs of Selective Catalytic Reduction of NOx. The taxed dispatch tends to use more units but at lower power. Additionally, the variance of expected emissions across the outage states of the system is reduced, reflecting the fact th at the dispatch optimization takes full account of the tax. Since during any one future episode the system state is random, the lower variance implies that the existence of the tax lowers the risk that NOx emissions will deviate from their expected value. Two power sector NOx emissions patterns, with and without the tax, were fed into BAAQMD’s version of the Urban Airshed Model, but the change in emissions resulted in an immeasurably small effect on peak ozone estimates for the September 1989 episode.