Nabila, Putri (2026) Monitoring Kualitas Air Produksi menggunakan Dynamic Principal Component Analysis (DPCA) dengan Hotelling's T2 dan Q-Statistic berbasis Bootstrap pada Perumda Air Minum Surya Sembada Kota Surabaya. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5003221029-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5003221029-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy |
Abstract
Air minum yang berkualitas dan memenuhi standar mutu menjadi kebutuhan masyarakat yang sangat penting untuk melindungi kesehatan dan mendukung kesejahteraan hidup. Oleh karena itu, kualitas air produksi perlu dipantau secara berkala karena perubahan pada satu variabel kualitas air dapat berkaitan dengan variabel lainnya dan juga dipengaruhi oleh kondisi pada periode sebelumnya. Penelitian ini bertujuan untuk menerapkan metode Dynamic Principal Component Analysis (DPCA) pada data kualitas air produksi Perumda Air Minum Surya Sembada Kota Surabaya serta menerapkan statistik Hotelling’s T² dan Q/SPE berbasis bootstrap mendeteksi penyimpangan proses. Variabel yang digunakan dalam penelitian ini adalah kekeruhan, pH, dan sisa chlor. Pengujian asumsi menunjukkan bahwa data tidak memenuhi asumsi normalitas multivariat, memiliki korelasi antarvariabel, dan autokorelasi, sehingga DPCA digunakan untuk menangkap struktur dinamis pada data. Diperoleh penentuan lag yang digunakan sebesar 4. Sehingga terbentuk matriks augmented dengan dimensi 257×15. Jumlah principal component yang dipertahankan adalah 14 dengan cumulative explained variance sebesar 97,27%. Batas kendali ditentukan menggunakan bootstrap dengan taraf signifikansi 0,00273. Pada fase I, pengamatan yang out of control dilakukan imputasi nol sehingga fase I merepresentasikan fase yang normal. Pada fase II, terdapat 20 dari 126 pengamatan atau 15,87% yang terdeteksi out-of-control berdasarkan statistik T², sedangkan berdasarkan Q/SPE terdapat 11 dari 126 pengamatan atau 8,73%. Penerapan DPCA berhasil menghasilkan ruang residual yang tidak menunjukkan autokorelasi signifikan, dengan p-value uji ljung-box Q-statistic pada fase I sebesar 0,8316 dan pada fase II sebesar 0,3935. Analisis kapabilitas menunjukkan bahwa secara multivariat proses produksi air minum telah kapabel dan mampu memenuhi batas spesifikasi Permenkes Nomor 2 Tahun 2023, yang ditunjukkan oleh nilai MC_np dan MC_npk lebih dari 1. Namun, kapabilitas tersebut belum merata pada seluruh variabel karena variabel Kekeruhan secara univariat masih belum kapabel dengan C_np sebesar 0,1993 dan nilai C_npk sebesar -0,2230.
====================================================================================================================================
High-quality drinking water is vital for protecting public health and supporting well-being. Therefore, treated water quality must be monitored regularly because its variables are interrelated and influenced by previous conditions. This study aims to apply the DPCA method to water quality data from PDAM Surya Sembada in Surabaya and to use bootstrap-based Hotelling’s T² and Q/SPE statistics to detect process deviations. The variables are turbidity, pH, and residual chlorine. Assumption testing showed that the data did not meet the assumption of multivariate normality, exhibited correlation among variables, and had autocorrelation; therefore, DPCA was used to capture the dynamic structure in the data. A lag of 4 was determined. This resulted in an augmented matrix with dimensions of 257×15. The number of principal components retained was 14, with a cumulative explained variance of 97.27%. Control limits were determined using bootstrap analysis with a significance level of 0.00273. In Phase I, observations that were out of control were imputed with zero, so that Phase I represents a normal phase. In Phase II, 20 out of 126 observations (15.87%) were detected as out-of-control based on the T² statistic, while based on the Q/SPE, 11 out of 126 observations (8.73%) were detected as out-of-control. The application of DPCA successfully produced a residual space that showed no significant autocorrelation, with a p-value for the Ljung-Box Q-statistic test of 0.8316 in Phase I and 0.3935 in Phase II. Capability analysis indicates that, multivariate analysis shows the drinking water production process is capable and meets the specification limits set by Permenkes No. 2 of 2023, as evidenced by MC_np dan MC_npk values greater than 1. However, this capability is not uniform across all variables, as the Turbidity variable is still not capable in univariate analysis, with a C_np value of 0.1993 and a C_npk value of -0.2230.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | Bootstrap, Dynamic Principal Component Analysis, Hotelling’s T2, Kualitas Air, Q-Statistic, Bootstrap, Dynamic Principal Component Analysis, Hotelling’s T2, Water Quality, Q-Statistic |
| Subjects: | H Social Sciences > HD Industries. Land use. Labor > HD9980.5 Service industries--Quality control. Q Science > QA Mathematics > QA278.5 Principal components analysis. Factor analysis. Correspondence analysis (Statistics) T Technology > TD Environmental technology. Sanitary engineering > TD259.2 Drinking water. Water quality |
| Divisions: | Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis |
| Depositing User: | Putri Nabila |
| Date Deposited: | 27 Jul 2026 01:23 |
| Last Modified: | 27 Jul 2026 01:23 |
| URI: | http://repository.its.ac.id/id/eprint/137759 |
Actions (login required)
 |
View Item |