Results and discussions Regarding turbidity, colour and COd in the secondary effluent from the wastewater treat- ment plant, we conducted water sampling and water quality analysis routinely over period of approximately I year. Taking turbidity as an example, a total number of 58 measurements ranging from 1.04 to 9.84 NTU were obtained, showing a wide spread ing of residual turbidity in the secondary effluent. By arranging all the measurements in an increasing order, calculating the cumulative probability, and plotting them on a log-normal coordinate system, Figure I was obtained. We further calculated the logarith- ×≤ mic average and standard deviation of these data as the characteristic parameters of the g-normal distribution In Figure 1, the dots are the measurement values and the curve is the fitted log-normal cumulative distribution using the calculated parameters. The result shows that the variation of residual turbidity in the secondary effluent follows the log-normal distribution well. It is therefore possible to use the log-normal distribution as a mathematical tool to carry out probabilistic analysis of the data. Similar analysis was conducted for residual colour and COd in the secondary effluent, as shown in Figure 2. Both the data of colour measurements(79 data) and COD measure- ments(10 data) fitted the log-normal distribution well a. Regarding total coliform and faecal coliform in the secondary effluent, the log-normal tribution was also applicable(Figure 3) Comparison of filtration processes for meeting the water quality targe After filtration, the concentrations of turbidity, colour and COD still followed the log-nor- mal distribution. In this study we compared three typical filtration processes, namely oagulation-filtration, O3-BAC, and UF as mentioned above. By log-normal probabilistic analysis of the water quality data, the effects of these filtration processes on turbidity colour and COD removal were evaluated Turbidity. Figure 4 shows the fitted probabilistic curves for turbidity residual in oagulation-filtration, O3-BAC, and UF treated water in comparison with the raw water i.e. the secondary effluent. Here, we took 5 NTU as the maximum allowable turbidity for urban wastewater reuse according to the Chinese standard (SAC, 2002). From the analysis result it was deduced that the probability for the secondary effiuent to meet 0.75 0.5 ◆ Measured 0.25 Turbidity/NTU Figure 1 Log-normal distribution of residual turbidity in the secondary effluentResults and discussions Pollutants distribution in the secondary effluent Regarding turbidity, colour and COD in the secondary effluent from the wastewater treat￾ment plant, we conducted water sampling and water quality analysis routinely over a period of approximately 1 year. Taking turbidity as an example, a total number of 58 measurements ranging from 1.04 to 9.84 NTU were obtained, showing a wide spread￾ing of residual turbidity in the secondary effluent. By arranging all the measurements in an increasing order, calculating the cumulative probability, and plotting them on a log-normal coordinate system, Figure 1 was obtained. We further calculated the logarith￾mic average and standard deviation of these data as the characteristic parameters of the log-normal distribution. In Figure 1, the dots are the measurement values and the curve is the fitted log-normal cumulative distribution using the calculated parameters. The result shows that the variation of residual turbidity in the secondary effluent follows the log-normal distribution well. It is therefore possible to use the log-normal distribution as a mathematical tool to carry out probabilistic analysis of the data. Similar analysis was conducted for residual colour and COD in the secondary effluent, as shown in Figure 2. Both the data of colour measurements (79 data) and COD measure￾ments (10 data) fitted the log-normal distribution well. Regarding total coliform and faecal coliform in the secondary effluent, the log-normal distribution was also applicable (Figure 3). Comparison of filtration processes for meeting the water quality target After filtration, the concentrations of turbidity, colour and COD still followed the log-nor￾mal distribution. In this study we compared three typical filtration processes, namely coagulation-filtration, O3-BAC, and UF as mentioned above. By log-normal probabilistic analysis of the water quality data, the effects of these filtration processes on turbidity, colour and COD removal were evaluated. Turbidity. Figure 4 shows the fitted probabilistic curves for turbidity residual in coagulation-filtration, O3-BAC, and UF treated water in comparison with the raw water, i.e. the secondary effluent. Here, we took 5 NTU as the maximum allowable turbidity for urban wastewater reuse according to the Chinese standard (SAC, 2002). From the analysis result it was deduced that the probability for the secondary effluent to meet 0 0.25 0.5 0.75 1 0.1 1 10 100 Turbidity/NTU Log-normal cumulative distribution Measured Calculated Figure 1 Log-normal distribution of residual turbidity in the secondary effluent X.C. Wang et al. 215