Total and Fractional Performance of Granular Dust Traps in Refractory Production

Abstract

Methods are proposed for designing interpolation models for the preliminary determination and subsequent forecasting of general and fractional breakthrough coefficients for dust used with granular filters, as employed in energy-saving and high-performance dust trapping from technological gases and ventilation discharges in refractory production. The models are supplied with nomograms, which makes them widely suitable for experts working in environmental protection at refractory-producing organizations. The main factors are identified that influence the performance. The results are of interest to experts in related areas of industry such as building materials and engineering ceramics and so on.The general and fractional breakthrough coefficients K and K j are major working characteristics of granular dust trapping filters in refractory production. The breakthrough coefficient isin which z i and z f are the dust contents of the gases or ventilation discharges before the trap and after it respectively in g/m 3 . The effects of actual conditions on the values of these quantities are so varied that it is impossible to present a unified and theoretically sound method of determining them. It is best to use interpolation models in the ranges for the actual physicochemical parameters of the flow and the geometrical characteristics of the trap. It has been found [1] that in generalin which w is the linear velocity of the dusty gas flow in m/sec; d e is the equivalent diameter of the pore channels in the filter layer in m; H is the thickness of the filter material in m; t is the filtration time in sec; d m is the mean median diameter of the dust particles in m; and s is the standard deviation of the logarithm for the particle diameters in a log-normal distribution (LND).Under those conditions, it is sound to plan an experiment to construct interpolation models by the Box-Wilson method with successive realization of short series of experiments on varying all the factors simultaneously. This enables one to approach the region of the optimum rapidly. In the experiments, the model dust was a polydisperse aerosol, which was used as two types of quartz dust with the following LND for the particle sizes: d m = 3.7´10 -6 m, s = 0.405, and d m = = 20´10 -6 m, s = 0.280; the factors d m and s are almost uncontrollable separately, so they are combined in a single control factor d m s. Table 1 gives the conditions, the planning matrix, and the results from the first series of experiments.In accordance with (2), we used the natural valuesx for the factors w, d e , Í, t, z i , d m s, which are denoted respectively byx 1 ,x 2 ,x 3 ,x 4 ,x 5 ,x 6 ; the encoded factors, variation levels, regression coefficients b i , and errors in determining them have been calculated by the method of [2]. Table 1 also gives the experimental values of the response function y = lnK -1 . The regression equation after checking the model for fit takes the form It follows from (3) that in these variation intervals for the factors, K increases with w and d e but ...