利用位移顯著性檢驗法進行水準點垂直變動偵測之研究

A Study of Test Method of Displacement Significance to Detect Vertical Movement of Benchmark

由於一等水準點位大部份皆沿著主要道路進行佈設及施測，若要形成閉合水準網型常需要很長之時間，其成果精度的保持往往只能以往返閉合差及環線閉合差來進行管制。本研究利用連續觀測3年之資料以形成3個環線包含173個一等水準點進行水準網整體平差，發現其中一個環線之觀測閉合差超過精度規範容許值，使得整體平差之點位中誤差最大達1.5cm。因此，利用內政部TWVD2001一等水準點資料與97年重測之一等水準點資料為檢驗依據，並利用平均間隙法及t檢驗法進行水準網整體與單點位移顯著性分析，檢驗該測量單位所觀測之一等水準點資料，找出差異過大之點位區域；續進行檢補測與重新計算並探討其閉合差超過規範容許值之成因。本研究檢測出兩段不良觀測量，將檢測觀測量加入重新進行整體平差計算後，環線閉合差已符合精度規範，且點位中誤差最大值已縮減為0.9cm，整體精度提升約40％。本研究採用之方法與流程，可有效分析誤差來源並縮小範圍，對後續水準測量相關工作具參考運用之價值。

The first order leveling survey and benchmarks are often done and set up along with the main road, so it takes a long time to form a closed networks. In this situation, the only way to make sure and keep precision is monitoring the differences of forward and backward measurements and circuit measurements. This study used 173 benchmark measurements in continued 3 years to create 3 circle loops, then did network adjustment and analyzed the result. The results showed the maximum of RMS is 1.5cm. Therefore, based on Taiwan Vertical Datum 2001 and the first order leveling data of 2008, we used The method of mean gap and t-test to analyze the displacement significance of the whole networks and the single point, examining the leveling data to find out the sinking area, then resurveying and discussing the reasons of the exceeding differences of circle measurements. In this study, two bad leveling sections have been detected. We added the leveling sections data into the adjustment, and the differences of circuit measurement had been in accordance with the precision, increasing 40% with RMS maximum decreasing to 0.9cm. This study demonstrates to analyze the sources of difference and to narrow the range, worthy to the leveling survey.