Yerel jeodezik GNSS CORS ağları ve bilimsel çalışmalara katkıları: İSKİ UKBS örneği

Abstract

İstanbul Su ve Kanalizasyon İdaresi (İSKİ), dünyadaki teknolojik gelişmeleri takip ederek ve kurumun ihtiyaçlarını göz önünde bulundurarak tüm İstanbul metropoliten alanını kapsayan ve ağ RTK (NRTK) prensibi ile çalışan Sürekli Çalışan Referans İstasyonları (Continuously Operating Reference Stations-CORS) ağını 2008 yılı sonunda kurmuş, ağ 'Uydulardan Konum Belirleme Sistemi (UKBS)' olarak isimlendirilmiştir. İSKİ UKBS Ağı ilk olarak 8 istasyon ile kurulmuş olup daha sonra ağın baz mesafeleri kısaltılarak geometrisi değiştirilmiştir. Mevcut durumda 10 istasyonu bulunmaktadır. Tez kapsamında, İSKİ UKBS Ağı'nın ülkemize ve bilime olan katkıları ifade edilerek, ağın sağladığı 10 yılı aşkın süredir devam eden statik veri ve düzeltme yayımı değerlendirilerek dört adet farklı senaryo ile uygulamalar gerçekleştirilmiştir. Birinci bölümde İSKİ UKBS Ağı hakkında genel bir bilgi verilerek, tezin amaçlarını ifade eden araştırma soruları, kullanılan yöntem ve materyaller ve literatür çalışması başlıklarına yer verilmiştir. İkinci bölümde İSKİ UKBS Ağı'nın kuruluş aşamasında gerçekleştirilen faaliyetler ele alınmıştır (Kurulum aşaması, istasyonların hassas konumlarının belirlenmesi, test yayım aşamaları, istatistik veriler). Ayrıca, İSKİ UKBS Ağı'nın kullanıcılara yayımlamış olduğu düzeltme modellerinin hesaplandığı iki sunucuda var olan, biri 2009-2015 yılları arası kullanılan ve şu an pasif olan GnSmart CORS yazılımı, diğeri 2015 yılından itibaren aktif olarak kullanılan Topnet Live CORS yazılımı incelenip analiz edilmiştir. Bu aşamada kullanıcıların sistemde kayıtlı tutulan logları, sistemin ürettiği düzeltme modelleri, çalışma saatleri ve ölçü kalitesi gibi farklı parametreler irdelenmiştir. Üçüncü bölümde, tezin amaçları doğrultusunda gerçekleştirilen, İSKİ UKBS Ağı düzeltme yayımı ve statik veri seti kullanılarak gerçekleştirilen dört farklı senaryoya ilişkin uygulamalar mevcuttur. Yapılan ilk uygulamada, istasyon hızları araştırılmıştır. Tektonik plaka ile uyumluluğu irdelenmiştir. İstasyon hızlarında yıllık trend dışında, mevsimsel, yarıyıllık ve yıllık salınım hareketinin varlığı araştırılmıştır. İstasyon hızlarının karşılaştırması yapılmıştır. Ayrıca, Global Navigasyon Uydu Sistemleri (Global Navigation Satellite System-GNSS) online değerlendirme servislerinin (Automatic Precise Positioning Service-APPS ve Australia Positioning Service-AUSPOS) konum belirleme performansı ve istasyon hızlarının belirlenmesinde kullanılabilirlikleri İSKİ UKBS Ağı istasyonları kullanılarak araştırılmıştır. İkinci çalışmada, İSKİ UKBS Ağı ve CORS-Türkiye (TR) Ağı baz mesafelerinin Sanal Referans İstasyonları (Virtual Reference Stations-VRS) çözümlerindeki belirsizlik çözüm süresi (Time to Fix Ambiguity-TTFA), tamsayı belirsizlik çözüm oranı (Fix rate), presizyon ve konum belirleme doğruluklarında ürettiği sonuçlar karşılaştırılmıştır. Üçüncü çalışmada, NRTK gözlemlerinde baz mesafesinin doğruluk ve presizyona etkisi araştırılmış ve yeni bir matematiksel model ile ifade edilmiştir. Ayrıca bu çalışma ile, Alan Düzeltme Parametreleri (Flaechen Korrektur Parameter-FKP) ve VRS yöntemlerinin farklı baz mesafelerindeki performansları analiz edilmiştir. Son çalışmada ise, İSKİ UKBS Ağı'nın Toplam Zenit Gecikmesi (Zenith Tropospheric Delay-ZTD) ve Bütünleşik Su Buharı (Integrated Water Vapour-IWV) kestiriminde kullanılabilirliği araştırılmıştır. ZTD ve IWV'nin lokal İSKİ UKBS Ağı'nda yükseklik ve mesafe farkına göre değişiminin incelendiği bu çalışmada, ayrıca IWV kestiriminde GNSS Analiz ve Konum Belirleme Yazılımı (GNSS Analysis and Positioning Software-GAPS) online servisinin performansı İSKİ UKBS Ağı istasyonları kullanılarak araştırılmıştır.

Istanbul Water and Sewerage Administration (ISKI) created the Continuously Operating Reference Station (CORS) network, which covers the whole Istanbul metropolitan area and works with the principle of the Network Real-Time Kinematic (NRTK), by following the technological developments in the world and considering the needs of the institution. ISKI CORS Network was first established with 8 stations and then the baseline length of the network was shortened and its geometry was changed. Currently, ISKI CORS Network continues its activities with 10 stations. Within the scope of the thesis, the contributions of ISKI CORS Network to our country and science have been expressed, and the static data and correction broadcast that has been carried out for more than 10 years has been evaluated and studies have been carried out with four different scenarios. In the first section of the dissertation, general information about ISKI CORS Network, the aim of the thesis, methods, materials and literature study are given. In the second section, the activities carried out during the establishment phase of the ISKI CORS Network are discussed (Planning, testing, installation, determining the precise positions of the stations and correction broadcast phase). Besides, the GnSmart CORS software, one of which was used between 2009-2015 and currently passive, and the Topnet Live CORS software, which has been used since 2015, was analysed. At this section, different parameters such as logs of the users, correction models produced by the system, working hours and measurement quality are examined. In the third chapter, there are studies related to four different scenarios realised for the thesis, using the ISKI CORS Network correction broadcast and static data set. In the first study, the station velocities were investigated. Its compatibility with the tectonic plate was examined. Except for the annual trend at station velocities, the existence of seasonal, semi-annual and annual oscillation movements was investigated. Station velocities were compared. Besides, the availability of Global Navigation Satellite System (GNSS) online process services (Automatic Precise Positioning Service: APPS and Australia Positioning Service: AUSPOS) in determining the positioning performance and station velocities were investigated using ISKI CORS Network stations. In practice, the academic BERNESE (v5.2) software, which can realise long-baseline length solutions for the true positioning technique, was selected and the results obtained with APPS and AUSPOS were analysed. The performance of the online services in positioning and velocity estimation was analysed with a statistical test at the level of α: 0.05 significance. In the study, local ISKI CORS Network stations were examined in 129 campaigns (one day per month) between 2008-2019 with BERNESE (v5.2) software. Also, the stations were examined in 34 campaigns using online services. As a result, the created velocity model showed high compliance with the sample data and NUVEL1A. Seasonal, semi-annual and annual oscillation movements were not detected. According to the statistical tests, APPS position values made a significant difference in the Z-axis. AUSPOS, on the other hand, gave statistically the same results as BERNESE (v5.2) results in positioning. In velocity accuracy, both services performed high; there was no significant difference between BERNESE (v5.2) results and online services. The subject of the geometric design of the CORS networks was investigated by using both local ISKI CORS Network and CORS-Turkey (TR) Network and redesign the ISKI CORS Network by increasing interstation distances. In this case study, interstation distances have been increased and the geometry of the network has been changed by using Topnet Live software, where the correction parameters in the control centre of ISKI CORS Network are calculated. ISKI CORS Network has been redesigned. In this way, the station frequency of ISKI CORS Network was analysed. Using three different CORS geometries (8-station ISKI CORS Network, redesigned 4-station ISKI CORS Network and CORS-TR Network), Virtual Reference Stations (VRS) solutions were obtained in real-time with RTKLIB software and tested in MATLAB with Analysis of Variance (ANOVA). As a result, the 8-station ISKI CORS Network with the smallest distance between stations in the vertical component gave the most accurate results. In TTFA performance, the CORS-TR Network performed worse than the other two CORS geometries. In fix solution rates, the redesigned ISKI CORS Network geometry solutions gave worse results than the other CORS geometries. It has been understood that the 8-station ISKI CORS Network performs better especially in vertical accuracy, also, TTFA performance and FIX solution rates are high, and its geometric design is correctly designed. In the third study, the effect of baseline length on accuracy and precision in NRTK measurements was investigated and expressed with a new experimental mathematical model. This study also provided an evaluation of the performance of Flaechen Korrektur Parameter (FKP) and VRS methods at different baseline lengths. Within the scope of this study, stations belonging to two CORS networks, one locally and one nationally, were used. Geodetic applications such as datum transformation between these two different networks and the transformation of the coordinate differences to the Local Geodetic System (LGS) were performed. In the analyses conducted, no significant change in Root Mean Square (RMSE) values was observed at baseline lengths of up to 40 km. On the contrary, it was determined that there was an average correlation coefficient of 0.665 between the precision and the baseline length linearly. To improve this correlation, a nonlinear variance model was created and tested as a function of the baseline length and the results were observed to be more compatible according to the linear model (the average correlation coefficient: 0.741). In addition, mean RMSE and precision values were examined in FKP and VRS measurements and no significant difference was observed. Average differences are in millimetre levels. In the last study, the usability of ISKI CORS Network in Zenith Tropospheric Delay (ZTD) and Integrated Water Vapour (IWV) estimation was investigated. While analysing the change of ZTD and IWV according to the difference in height and distance in the local ISKI CORS Network, the performance of the GNSS Analysis and Positioning Software (GAPS) online service was investigated using ISKI CORS Network stations. For the ZTD estimation, GnSmart software, which broadcasts CORS correction was used, and GAPS was used for the IWV estimation. Within the scope of the study, the effects of height differences and distances on ZTD and IWV estimation were analysed to investigate the usability of ISKI CORS Network stations in weather prediction studies. Accordingly, it was observed that the height differences for ISKI CORS stations affected ZTD more (Correlation Coefficient: 0.95). However, when we look at the annual averages, there was no difference for ZTD in seasonal movement between stations. Radiosonde station 17064 data inside the network was used as a reference for the IWV analysis. Accordingly, RMSE values in the range of ± 3.4mm / ± 7.6mm and correlation coefficients in the range of 0.69-0.92 were calculated. It is understood that distance differences affect IWV more than height difference. With this study, it was stated that in addition to meteorological observations, ZTD and IWV values obtained from ISKI CORS Network stations can be used in weather forecasting studies and it is stated that GAPS Online Service can be used for IWV estimation.