<![CDATA[Background: Step-shaped cascades are recognized as the most effective in reducing Fe in water compared to flat or wavy cascades. However, previous studies have only been limited to identifying the optimal cascade shape and have not yet reached the development of a predictive model for Fe reduction using this cascade method.Objective: Developing a predictive model for the reduction of Fe in water using a cascade made of layered wood coated with plastic/rubber, with water flow diameters using hoses/pipes of 3/16 inch, 5/16 inch, and 7/16 inch.Method: The type of study is an experiment, using 75 water samples taken from dug wells or hand pump wells, which were channeled through the cascade. The dependent variable was the reduction of Fe after aeration through the cascade, while the independent variables were water flow rate, water pH, water temperature, dissolved oxygen, and Fe content in the water before aeration. The data were analyzed using multiple linear regression. This study did not measure or observe humidity, air temperature, and oxygen content in the air because it was conducted in an environment with consistent conditions.Results: A candidate model was obtained consisting of four main variables, i.e. water flow rate 1 and water flow rate 2, water pH, dissolved oxygen, and initial Fe. There were two interaction variables, i.e. the interaction between water flow rate and Fe before aeration, and the interaction between dissolved oxygen and Fe before aeration. Diagnostic results regarding assumption fulfillment indicated that the candidate model satisfies the assumptions of multiple linear regression; however, collinearity diagnostics revealed the presence of collinearity symptoms within the model. However, since the model was intended for prediction, therefore collinearity is not a serious issue. The model reliability test yielded a Shrinkage value of 0.21 or 21%, which means that the candidate model is sufficiently reliable to be established as a predictive model for the reduction of Fe in water using the cascade method.Conclusion: The mathematical function equation obtained for Fe reduction as the final result of the study is: -695 + (-0.166 × flow rate 3.7 l/min) + (-0.334 × (flow rate 1.06 l/min) + 0.329 × (water pH) + (-0.108 × dissolved oxygen)) + (-0.664 × initial Fe)) + 0.142 × (flow rate 1 × initial Fe) + 0.588 × (flow rate 2 × initial Fe) + 0.134 × (dissolved oxygen × initial Fe). Keywords: prediction model, Fe reduction, cascade method ABSTRAK Latar Belakang : Cascade bentuk anak tangga diketahui sebagai yang terbaik dalam menurunkan Fe dalam air dibandingkan bentuk Cascade datar atau bergelombang, namun penelitian terdahulu hanya terbatas mengetahui bentuk cascade terbaik, tidak sampai memperoleh model prediksi penuruan Fe dalam air menggunakan metode cascade.Tujuan : Tujuan penelitian ini adalah dapat menghasilkan model prediksi penurunan Fe dalam air menggunakan cascade dengan bahan kayu berlapis plastik/karet dengan diameter aliran air menggunakan selang/pipa 3/16 inchi, 5/16 inchi dan 7/16 inchi.Metodologi : Jenis Penelitian adalah eksperimen dengan menggunakan 75 buah sampel air yang berasal dari sumur gali atau sumur pompa tangan kemudian dialirkan melalui cascade. Variabel Dependen adalah penurunan Fe setelah aerasi melalui Cascade dan Variabel Independen adalah Debit air, Derajat Keasaman Air, Suhu Air, Oksigen Terlarut dan Kandungan Fe dalam Air sebelum aerasi, kemudian dianalisis menggunakan Regresi Linier Ganda. Didalam penelitian ini tidak melakukan pengukuran/ pengamatan terhadap kelembaban, suhu udara dan kandungan oksigen di udara karena dilakukan pada ruang dengan kondisi yang sama. Hasil : Dari hasil penelitian diperoleh kandidat model yang berisi empat variabel dasar yaitu variabel Debit 1 dan debit 2, Derajat Keasaman, Oksigen Terlarut dan Fe Awal serta dua variabel interaksi yaitu interaksi antara Debit Air dan Fe sebelum aerasi serta Interaksi antara Oksigen Terlarut dengan Fe sebelum aerasi. Hasil diagnostik pemenuhan asumsi menyatakan bahwa kandidat model dapat memenuhi asumsi-asumsi regresi linier ganda namun dari diagnostik kolinearitas diketahui terdapat gejala kolinearitas dalam model, tetapi karena model bertujuan untuk prediksi maka masalah kolineraitas bukan merupakan masalah serius. Uji reliabilitas Model memberikan hasil nilai Shrinkage = 0,21 atau = 21 %, dengan demikian kandidat model cukup reliabel untuk ditetapkan sebagai model prediksi penurunan Fe Dalam Air menggunakan Metode cascade.Kesimpulan : Persamaan fungsi matematis Penurunan Fe yang diperoleh sebagai hasil akhir dari penelitian adalah:Penurunan Fe = - 695 + (- 0,166 x debit 1) + (- 0,334 x (Debit 2) + 0,329 (pH Air) + (-0,108 x Oksigen Terlarut)) + (-0,664 x Fe Awal)) + 0,142 x (Debit1*Fe Awal) + 0,588 x (Debit 2 * Fe Awal) + 0,134 x (Oksigen Terlarut * Fe Awal). Kata Kunci : Model Prediksi, Penurunan Fe, Cascade]]>
Copyrights © 2026
