Optimization of a particulate control train design using a log-bimodal distribution.

摘要

A methodology was developed for the determination of the optimal design of a control train to reduce emissions of particulate matter with ultra high efficiency (99.99%) when the inlet particle size distribution is log bimodal. The control train optimized consisted of a multiple cyclone pre-cleaner followed by a wire-plate electrostatic precipitator (ESP). The objective function for optimization was minimization of the total annual cost (TAC) for the combined control train. The optimization variable studied was the total mass penetration through the cyclonic pre-cleaner. Brent's method for unidimensional optimization without derivatives was used to isolate the optimal design. Marquardt's method was used to estimate, by nonlinear regression analysis, the five parameters of the log bimodal particle size distribution at the pre-cleaner exit. The methodology developed was applied to a case study: control of particulate emissions from the waste-to-energy conversion system of a paper mill.; Another factor considered in the optimization of the case study was the effect of back corona in the ESP. Back corona results when high resistivity dust is collected in the ESP. The presence of back corona reduces the intensity of the electric field with significant deterioration of the ESP performance.; It was found that the optimal design with no back corona was a multiple cyclone system consisting of 2,895 Stairmand cyclones of 0.20 m external diameter with a total mass efficiency of 84%, followed by an ESP with a total collection area of 26,300 {dollar}rm msp2.{dollar} The corresponding TAC was {dollar}2.48 millions/yr. When moderate back corona was considered, the number of cyclones increased to 3,102; the ESP collection area increased to 75,900 {dollar}m sp2;{dollar} and the TAC increased to {dollar}4.01 millions/yr.