MIN Byung Dae, Graduate School of Urban Environmental Sciences, Department of Civil and Environmental Engineering
| In small-scale water systems, measurement of the quality of raw water for water supplies is generally conducted under normal conditions when the quality of water is stable. Moreover, since there are only slight variations in the water quality, measurements are often taken infrequently. However, characteristics such as water temperature and pH, which are easily monitored, are regularly recorded. In establishing an improvement plan for water treatment systems, the range of concentration of the target material present in the raw water for water supplies provides useful and valuable information. If the concentration of the target material can be specified by the quality of water through the measurements that are collected daily, estimation of the range of concentration in the past is feasible as well. |
| The aim of research is to estimate over the past five years the concentration levels of manganese in raw water for water supplies at the Ohgiura water treatment plant in Ogasawara- village, Tokyo. This facility provides a small-scale water system based on estimation results and establishes the most appropriate manganese treatment system for the target area. |
|
Explanatory variables used for the estimation equation were the same as the data measurement characteristics (water temperature, turbidity, pH, color, alkali level, and electrical conductivity), and by assigning daily data to the estimation equation obtained by multiple regression analysis, daily data of manganese concentrations for the past five years were estimated. |
|
|
Fig.1.Monthly data of water quality measurement item of raw water manganese. |
|
|
Fig.2.Monthly data of manganese |
Table.1.Result of correlation analysis.
Explanatory
variables
|
Relative coefficient with
manganese |
| Not converted |
After logarithmic conversion |
| Water temperature |
0.288 |
0.332 |
| Turbidity |
0.374 |
0.422 |
| color |
0.197 |
0.370 |
| ｐH |
-0.109 |
-0.051 |
| Alkali level |
-0.061 |
-0.097 |
| Electrical conductivity |
-0.096 |
-0.230 |
| （n=57, r99=0.335） |
|
| Correlation analysis was implemented for six characteristics used in estimating the concentration of manganese. Correlation analysis was also provided for logarithmically converted data, taking into account nonlinear relationships. |
|
|
Fig.3.Estimation of daily data of manganese concentration |
| As for turbidity, which has the highest correlation with manganese concentration, averages of monthly data and daily data were almost identical. However, the ratio of the maximum value and the average value were 2.6 in the monthly data used for the derivation of the estimation equation, but 11.0 in the daily data. This demonstrates that the measurement of manganese concentration was not collected when turbidity was high. When 11.0 (ratio of the maximum value and the average daily turbidity data) is multiplied by the average of manganese concentration estimated by daily data, 0.89 mg/L is derived. |
| The maximum value of 1.02 mg/L at the 95% confidence interval was used as the upper limit of the concentration of manganese in raw water for water supplies. Based on this estimation the authors have proposed a water treatment system for manganese removal. |
returns to contents