| Summary: | 碩士 === 中華大學 === 土木工程學系碩士班 === 100 === Whether the water is polluted or not, it can be identified by its salinity or alkalinity. For the salt damage, the electrical conductivity represents the dissoluble inorganic salts in water. For the alkali damage, the Sodium Adsorption Ratio (SAR) is used to identify the classification of irrigation water quality and to know the harm and range of the pollution using the graphical classification of relation between electrical conductivity and SAR in the irrigation water quality management guide. This study adopted the historical water examination data provided by Hsinchu Irrigation Association, and used water temperature (T), pH, electrical conductivity (EC), Ca2+ and Na+ as input factors. The regression analysis and the genetic algorithm combined with operation tree (GAOT) were used and compared. The data were divided into 25 data of training simulation and 13 data of test simulation, and the relation scheme was built. The linearly correlation coefficient (CC) value, Root Mean Squared Error (RMSE) and the relevance between the actual value of SAR and the predicted value of SAR were used to check the quality of analytic methods. Finally, according to the scatter diagram result, the predictability of GAOT analysis is better than that of regression analysis, and it is feasible to use it and real-time monitoring data of water quality to estimate the salt and alkali damage. It can replace traditional manual on-site sampling for examination, which is labor and time consuming.
|
|---|
