EVOLUTION OF RESIDUAL CAKE HEIGHT, CAKE PATCHES, AND REGENERATION EFFICIENCY ON NEEDLE FELT FILTER MEDIA

Transcription

1 EVOLUTION OF RESIDUAL CAKE HEIGHT, CAKE PATCHES, AND REGENERATION EFFICIENCY ON NEEDLE FELT FILTER MEDIA *Mahmood Saleem, Institute of Chemical Engineering & Technology, University of the Punjab, Lahore, Pakistan [This work was done at Graz University of Technology, Graz, Austria] *Rafi Ullah Khan, Institute of Chemical Engineering & Technology, University of the Punjab, Lahore, Pakistan ***M. Suleman Tahir, Graz University of Technology, Graz-Austria ***CorrespondenceAuthor: Abstract: Residual filter cake onfilterbags is responsible for the residual pressure drop and evolution of the pressure drop during the following filtration cycle. Distribution of residual cake depends on the efficiency of regeneration which further depends on many factors like dust and bag filter characteristics, operating conditions, and the pulse jet cleaning. This paper investigates the evolution of residual cake height, resulting cake patches area distribution, and local regeneration efficiency on pulse jet cleaned needle felt filter bags in a pilot scale bag filter facility. Results reveal an influence of location on evolution of cake height distribution, cake area distribution, and regeneration efficiency. Keywords: Residual Filter, Efficiency, Regeneration, Evolution 1. Introduction Bag filters are frequently employed in process industry for fine solids removal from process gases. Their behaviour is a complex phenomena [1]. It is importantto consider area distribution of cake height as well as local regeneration efficiency for the prediction of filter behaviour overtime [2]. Based on variations in surface behaviour or other reasons, some areas of the bag may have lesser tendency of regeneration. Therefore, cake formed on such sections may survive many cleaning cycles and eventually become un-detachable by the jet pulses which might have consequences on filter operation [3]. Many attempts are reported [4] [5] [6] [7] [8]. This paper presents the experimental investigations of the evolution of residual cake height, resulting cake patches area distribution, and local regeneration efficiency on pulse jet cleaned needle felt filter bags in a pilot scale bag filter facility. Test is also performed to investigate filter cake survival from cleaning cycles. 2. Experimental Conditions The study is carried out in a filtration test where limestone dust (d 503 =5 m) is being separated from air at ambient conditions. The filter media is Polyphenylenesulfide needle felt bags. The mean air velocity is 27.3 m/s at inlet dust concentration of 4.29 g/m 3. The bags are jet pulse cleaned at 2.5 bar. The impulse time is 20ms and dispersion air is at 5 bar. Relative humidity is recorded as 45% at 25 C. Initially a ramp test is carried out for the determination of resistance parameters with already conditioned filter bags. Then, filtration test is carried out. There are three bags, one in each row. One bag is subjected to jet pulses at the upper limit of pressure drop. Optical measurement is taken for cake height after regeneration at 7 points in the test. The data are evaluated for residual cake height distribution evolution, residual cake patches area distribution, and frequency of regeneration using procedures presented elsewhere [9]. 3. Regeneration Efficiency Cake height images are evaluated using the following criteria: Let i be the number of pixels in 1 st dimension and j be the no. of pixels in 2 nd dimension. Thus for all i andj If, h(ij) =100m H(i,j)=l else h(ij)=0 here h(i,j) represents cake height corresponding to pixels in the image. For N measurements, the sum of arrays will have values between 0 and N. Divided by N will convert values between 0 and 1. Thus a region never regenerated will have a value of zero. Thus the local regeneration efficiency is obtained.

2 4. Results and Discussions 4.1 Cake height evolution The cake height measurement corresponding to 10%, 50%, and 90% cumulative area are presented in Figure 1, Figure 2, and Figure 3 at top, middle, and bottom of the measured section respectively. The numbers on abscissa refer to Table 1. Table 1: Residual pressure drop The difference between X 0 9 and X 0A might be used as an indicator of cake height distribution. The median cake height X 05 is also included. The difference between the first two is nearly 0.3mm, 0.25mm, and 0.5mm at the top, the middle, and the bottom respectively. These measurements indicate that distributions are relatively wider at the bottom and that the median cake height at the bottom is the highest. If one looks at each of the measurements, the scatter in the measurements might be attributed to AP [Pa] Ref. No Filtration Cycles the scatter of residual pressure drop. The variation in the trends at three locations tempts to believe that regeneration is influenced by the location resulting in variations of regeneration behaviour. I Fig. 1: Evolution of Cake height corresponding to 10%, 50%, and 90% of cumulative area at the top of the measured section. i o e 0.1 a 0 '5 J= -0.1 w O & - & e X B O- x Fig. 2: Evolution of Cake height corresponding to 10%, 50%, and 90% of cumulative area at the middle of the measured section Fig. 3: Evolution of Cake height corresponding to 10%, 50%, and 90% of cumulative area at the bottom of the measured section 4.2 Evolution of Area Distribution of Residual Cake Patches Assuming the cake detachment a random process, the cake patches area distributions should be random as well. Residual cake patches area distributions are presented in Figure 4, Figure 5, and Figure 6 at three locations on the measured section at the top, middle, and bottom respectively. It is obvious that smaller patches appear many while large patches appear very few. However, distributions change across the measurements indicating their random nature. One can observe that distributions at the bottom section

3 are significantly different than those at the middle, or the top sections. This reflects upon its location dependency. Fig. 4: Residual cake area distributions at the top section Q 60 r 1 e O 16 - a T = C mm Fig. 5: Residual cake area distributions at the middle section Q 60 r 1 e O 16 - a C mm Fig. 6: Residual cake area distributions at the bottom section. 4.3 Regeneration Efficiency of Filter Surface Before discussing regeneration efficiency, it is better to look at the geometric constraints of the filter bag as shown in Figure 7. Grayscale images of regeneration efficiency at the top, middle, and bottom of the measured section are displayed in Figure 8. Some area is dark black and it is not regenerated at all. Rest of the area shows regeneration efficiency between 30-40%. At the top right of (a), bottom right of (b), and in the middle of (c) are regions which are not regenerated. There exists a residual cake layer thicker than 100m. If one considers the relative position of the measured spots, nearly 2cm from the support ring measurement starts. Top section (a) is 6cm away from the circular rod. The middle section (b) is far away from the circular rod. However, the bottom section (c) is close to the ring. Measured sections are between longitudinal rods. Therefore, one can assume that the bottom section, close to support ring, does not bend inward because of increasing pressure drop due to dust cake build up. A low regeneration at this region seems to be linked to the presence of cage rod behind the measuring section. It is reasonable to believe different cake detachment behaviour at different locations on the bag. However, it is not fully resolved whether this behaviour is solely because of the cage rods. 5. Conclusions Optical cake height measurement technique is employed to examine the residual distribution of cake height, cake patched, and regeneration of filter media is a pilot scale jet pulsed bag filter facility where limestone dust is being separated from air on polyphenylenesulfide needle felt bags at ambient conditions. Optical measurements of cake height distributions indicate varying trends at different locations which tempts to believe that regeneration is influenced by the location resulting in variations of regeneration behaviour. In-line with cake height distributions, cake patches area distributions change across the measurements indicating their random nature. This reflects upon its location dependency.

4 30mm the financial support of FWF (Austria) and HEC (Pakistan) for this work. Supply of needle felt by M/S Inspec Fibers (Lenzing, Austria) is also acknowledged. References Regeneration efficiency is measured at 7 points in the test. The analysis reveals different cake detachment behaviour at different locations on the bag. However, it is not fully resolved whether this behaviour is solely because of the cage rods. Acknowledgments This work was performed at Graz University of Technology, Graz, Austria. Authors acknowledge [1] F. Loefler, H. Dietrich, and W. Flatt, Dust Collection with Bag and Envelop Filters, Fried Vieweg and Sons, Braunschweig, Germany, [2] A., Dittler and G., Kasper, Simulation of operational behaviour of patchily regenerated, rigid gas cleaning filter media, Chemical Engineering and Processing, Vol., 38, 1999, pp: [3] A. Kavouras, G. Krammer, Distributions of age, thickness and gas velocity in the cake of jet pulsed filters application and validation of a generations filter model, Chemical Engineering Science, Volume 58, Issue 1,

5 January 2003,pp: [4] Stefan Berbner, Torsten Pilz, Characterization of the filtration and regeneration behaviour of rigid ceramic barrier filters at high temperatures, Powder Technology, Volume 86, Issue 1, January 1996, pp: [5] A. Dittler, H. F. Umhauer, The influence of conditioning and regeneration on the separation behaviour of rigid surface filters for the separation of particles from gases, Powder Technology, Volume 120, Issue 3, 22 October 2001, pp: [6] A. Dittler, M. V. Ferer, P.t Mathur, P. Djuranovic, G. Kasper, D. H. Smith, Patchy cleaning of rigid gas filters transient regeneration phenomena comparison of modelling to experiment, Powder Technology, Volume 124, Issues 1-2, 8 April 2002, pp: [7] M. Bogdanic, F. Behrendt, F. Mertins, The influence of a 2-component model on the computed regeneration behaviour of an uncoated diesel particulate filter, Chemical Engineering Science, Volume 63, Issue 10, May 2008, Pages [8] Ch. Stocklmayer, W. Hoflinger, Simulation of the regeneration of dust filters, Mathematics and Computers in Simulation, Volume 46, Issues 5-6,1 June 1998, Pages [9] M. Saleem, Gas Cleaning: An Experimental Study, VDM Verlag, Germany, 2010, ISBN= ,