Sensitivity analysis for a proposed sewer overflow screening device

Sewer overflows to receiving water bodies cause serious concerns for the environment, aesthetics and public health. To overcome these problems a self-cleansing, low maintenance, high capture efficiency and less expensive device was developed and tested at Swinburne University of Technology, Melbourne. There are a number of different screening systems used in sewer overflow screening devices. Most of the screener has the common drawbacks in the available commercial devices include inadequate screening capacity, external power needs and high cost. To overcome such drawbacks a new overflow sewer device, known as the 'Comb Separator' was proposed. The device has no moving parts, a robust stop/start operation, an effective self-cleansing mechanism, low maintenance and operation costs and no external power requirements. The proposed experimental device in the present research is sewer gross pollutant trapping device, which consists of a rectangular tank and a sharp crested weir. In front of the weir a series of vertical, parallel combs to separate entrained sewer solids from the overflow as shown in Figure 1. The studied device was tested with a series of sewer solid materials including condoms, tampons, cigarette butts, cotton buds, bottle caps, wrap papers etc. Larger sewer particles (greater than 10mm diameter) can be captured relatively easily with capture efficiency more than 90%. This capture efficiency was tested with different input varying condition; however the output capture efficiency has insignificant variation from varying input parameters. However, significant variation observed from varying input parameters for smaller particles. Hence focus of this work is to parameter sensitivity on smaller sewer solid particles. To improve output capture efficiency of these smaller particles, important input parameters like flow conditions, layers of combs and spacing of combs and weir opening were changed and tested with different trials. Average capture efficiency varies from 50% to 85%, however at times capture efficiency varied without varying input parameters, which triggered the need for a detail investigation of the parameter sensitivity. A total of forty (40) sets of experimental data were collected on eight different sets of experimental setup. Based on the experimental experience, four input parameters (flow volume, effective combs spacing, weir opening and number of comb layers) were identified influential on output sewer capture efficiency. Four different method of sensitivity testing were adopted these are Partial plots, Partial Correlation Coefficient (PCC), Sensitivity Index (SI) and Regression Analysis. The partial plots suggest that inflow volume have a negative correlation with the output sewer capture efficiency. The 1st comb spacing has a positive correlation whereas the 2nd comb separator has a negative correlation. The weir opening has a positive correlation as wider weir area will reduce velocity and increase capture efficiency. Partial correlations suggest that weir opening is the most important parameter followed by 2nd comb spacing, inflow volume and 1st combs spacing. Sensitivity Index suggests all four parameters are varied 5% only for their sensitivity index. Regression analysis did not consider weir opening and comb layers as input parameters since these parameters failed to satisfy normal distribution assumption. Effect of 1st and 2nd comb spacing are combined in effective comb spacing (which shows bell shape trend of the input parameters) along with inflow volume as input parameters. Results shows inflow volume has significant influence output capture efficiency of the output parameter. Effective comb spacing combine the effect of comb does not provide any particular trend over capture efficiency although 1st comb has a positive correlation and 2nd comb spacing has a negative correlation with the capture efficiency.