AAAR 31st Annual Conference
October 8-12, 2012
Hyatt Regency Minneapolis
Minneapolis, Minnesota, USA
Abstract View
Development of a New SMP Model Satisfying All Known Physical Constraints in Aerosol Source Apportionment Study
BONG MANN KIM, AQMD
Abstract Number: 632 Working Group: Source Apportionment
Abstract A new multivariate receptor model, SMP (Solver for Mixture Problem), has been developed that satisfies all known fundamental natural physical constraints identified in aerosol source apportionment study. Previously, Nonnegative Matrix Factorization (NMF) models have been developed and applied successfully in chemometrics and PM source apportionment study. NMF/PMF models were developed based on nonnegativity constraints of loading and scores (equivalent to source compositions and source apportionments in receptor modeling). However, more physical constraints need to be satisfied in addition to nonnegativity constraints of source compositions and source apportionments. A new model has been developed based on alternating primal-dual interior point nonlinear programming, subject to the inequality constraints of all known physical constraints. Although many multivariate receptor models attempt to estimate both source compositions and source apportionments from the ambient data alone, the SMP model is the first model that can estimate both source apportionments and source compositions. Source compositions estimated by the SMP model can be used in other CMB analysis. The SMP model has been applied to an error free data set to examine its capability of estimating source compositions and source apportionments. Three simulations were conducted, and the results will be compared and discussed.