Please use this identifier to cite or link to this item:
https://ptsldigital.ukm.my/jspui/handle/123456789/464294Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor.advisor | Wayan Suparta, Prof. Madya Dr. | |
| dc.contributor.author | Kemal Maulana Alhasa (P67736) | |
| dc.date.accessioned | 2023-09-26T09:26:12Z | - |
| dc.date.available | 2023-09-26T09:26:12Z | - |
| dc.date.issued | 2015-04-13 | |
| dc.identifier.other | ukmvital:81639 | |
| dc.identifier.uri | https://ptsldigital.ukm.my/jspui/handle/123456789/464294 | - |
| dc.description | Wap air adalah salah satu komponen dalam atmosfera Bumi yang berperanan penting dalam peramalan cuaca, terutamanya dalam bidang peramalan hujan, mitigasi bencana serta pengurusan air. Sistem Pendudukan Global (Global Positioning System-GPS) telah digunapasti sebagai sebuah alat berkos rendah dalam penganggaran wap air. Pun begitu, prestasi sistem GPS ini didapati masih kurang menyediakan data yang berterusan selama 24 jam penuh. Ini boleh menjejaskan kepada kesediaan data taburan wap air dan seterusnya ia boleh memberi kesan kepada ketepatan sistem peramalan cuaca. Oleh itu, objektif utama bagi kajian ini ialah untuk membina sebuah sistem alternatif yang boleh memantau dan memberi maklumat wap air boleh mendak (Precipitable Water Vapor–PWV) tanpa perlu menggunakan sistem penerima GPS. Untuk mencapai objektif ini, kaedah pendekatan sistem inferens kabur neural adaptif (Adaptive Neuro Fuzzy Inference System ANFIS) digunakan untuk membina model anggaran jumlah lengahan zenit (Zenit Total Delay-ZTD) dan PWV dengan tiga input data meteorologi permukaan iaitu tekanan (P), suhu (T) dan kelembapan relatif (H). Data meteorologi permukaan daripada lima stesen terpilih di kawasan Antartika dan ZTD yang diperolehi daripada CDDIS NASA bermula dari Januari hingga Disember 2010 digunakan sebagai data latihan dan pengesahan bagi model ZTD di kawasan Antartika. Manakala data meteorologi permukaan dan, ZTD serta PWV daripada ukuran GPS bumi di tiga stesen terpilih kawasan semenanjung Malaysia dan Sabah bermula dari Januari hingga Disember 2009 digunakan sebagai data latihan dan pengesahan bagi kedua-dua model ZTD dan PWV di kawasan khatulistiwa. Seterusnya, kaedah rangkaian neural tiruan (Artificial Neural Network-ANN) serta regresi linear berganda (Multiple Linear Regression–MLR) digunapakai sebagai pembanding kebolehupayaan model ANFIS yang dibangunkan. Sistem ini dibangunkan dengan menggunakan perisian LabView. Data hasil pencerapan meteorologi dan penganggaran nilai ZTD dan PWV daripada model ANFIS disimpan ke dalam sistem pengurusan pangkalan data untuk dipaparkan ke dalam laman sesawang secara berterusan. Hasil kajian menunjukkan bahawa model ZTD dan PWV bersama ANFIS dengan tiga input meteorologi (P,T,H) berkorelasi positif dan berpadanan dengan nilai ZTD yang diperoleh daripada CDDIS NASA dan TroWav, serta PWV yang diterbitkan daripada sistem GPS. Keputusan daripada subset pengesahan menunjukkan bahawa model ANFIS mempunyai ralat yang paling rendah berbanding dengan model ANN dan MLR, dengan peratus ralat 0.034%, min ralat mutlak 0.015 mm dan punca min ralat kuasa dua 0.021 mm. Hasil ini memberi keyakinan bahawa model ANFIS dengan masukan meteorologi yang telah dibangunkan boleh digunapakai sebagai sistem pemantauan masa-nyata untuk menganggarkan ZTD dan PWV. Sistem ini telah diuji dan mempunyai keupayaan untuk memapar maklumat semasa dalam bentuk graf dan papan tolok. Di samping itu, ia boleh menyimpan fail secara automatik ke dalam komputer dan sitem pangkalan data. Walau bagaimanapun, ketepatan sistem ini bergantung kepada kualiti data latihan yang diberikan kepada model ANFIS semasa proses latihan. Dengan sistem yang telah dibangunkan di UKMB, nilai ZTD dan PWV boleh terus dianggarkan tanpa memerlukan penerima GPS dan maklumat yang dihasilkan ini boleh diaplikasikan untuk kajian meteorologi.,Water vapor is one of the components in the Earth's atmosphere, which has an important role in weather forecasts, especially in the field of rainfall forecasts, hazard mitigation, and water management. Global Positioning System (GPS) has been used as a low-cost tool for estimation of atmospheric water vapor. However, the performances of GPS system are still lacking of providing continuous data for a full 24-h period. This problem can affect the continuous availability of water vapor data and accuracy of weather forecast system. Thus, the main objective of this study is to build an alternative system, which can monitor and provide information about the precipitable water vapor (PWV) without using a GPS receiver. To fulfill this objective, the adaptive neuro fuzzy inference system (ANFIS) approach is used to develop an estimation model of Zenith Total Delay (ZTD) and PWV based on the three surface meteorological data as inputs (pressure (P), temperature (T) and relative humidity (H). A large amount of the surface meteorological data from five stations selected in the Antarctic region and ZTD derived from CDDIS NASA from January to December 2010 are used to train and validate process of ZTD model in Antarctic region. While the surface meteorological data and, ZTD and PWV from GPS measurements at three stations selected in peninsula Malaysia and Sabah region from January to December 2009 are used for the training and validation process of both models ZTD and PWV in equatorial region, respectively. In addition, the artificial neural network (ANN) and multiple linear regression (MLR) methods were employed as a standard of comparison to the ANFIS model. This system was developed using the LabView programming. Data from the surface meteorological measurement as well as ZTD and PWV estimation from the ANFIS model is stored into database management system and displayed on the website. The results showed that both ZTD and PWV ANFIS with three meteorological inputs (P, T, H) were positively correlated and comparable with the ZTD values from CDDIS NASA and TroWav, and PWV derived from GPS observation. In addition, the result from the validation subset showed that the ANFIS model has the lowest error compared to ANN and MLR models, with a percent error of 0.034%, a mean absolute error of 0.015 mm and a root mean square error of 0.021 mm. This gives confidence that the ANFIS model developed with the input of surface meteorology can be applied into the real-time monitoring system to estimate ZTD and PWV. This system has been tested and has the ability to display the current information in terms of graph and gauge board. More than that, the system has the ability to store files automatically into the computer and database systems. Overall, the accuracy of the system is depended on the training data that given into ANFIS model during the training process. By the system that was developed at the UKMB station, the ZTD and PWV can be estimated directly without the GPS receiver, and the information obtained can be applied for meteorological studies.,Sarjana | |
| dc.language.iso | may | |
| dc.publisher | UKM, Bangi | |
| dc.relation | Institute of Climate Change / Institut Perubahan Iklim | |
| dc.rights | UKM | |
| dc.subject | Wap air | |
| dc.subject | Sistem inferens kabur neural adaptif | |
| dc.subject | Sistem peramalan cuaca | |
| dc.subject | Water vapor | |
| dc.subject | Atmospheric. | |
| dc.title | Pembangunan sistem pemantauan masa-nyata wap air atmosfera menggunakan sistem inferens kabur neural adaptif | |
| dc.type | theses | |
| dc.format.pages | 199 | |
| dc.identifier.callno | QC915 .K444 2015 | |
| dc.identifier.barcode | 001443 | |
| Appears in Collections: | Institute of Climate Change / Institut Perubahan Iklim | |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| ukmvital_81639+SOURCE1+SOURCE1.0.PDF Restricted Access | 18.97 MB | Adobe PDF |  View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.