建構以實驗室法進行慣性式平坦儀認證之程序與資料分析方法

Abstract

對慣性式平坦儀之設備認證與準確性之研究，大多利用已知剖面之實測道路進行，此評估方式包含設備硬體(加速度規、距離感測器、位移感測器)、分析軟體(訊號處理、指標計算)及操作人員穩定性(重複行駛速率與路徑控制)之綜合誤差，且在認證進行過程中，駕駛者無法維持同一行駛路徑，便難以進行量測高程準確性比對，若採用慣性式平坦儀作為施工驗收或鋪面養護管理決策判定基準，應避免人員操作誤差所產生之影響，需有效將設備硬體與分析軟體之評估獨立出來，以確實掌握量測設備之精準度。為獨立進行慣性式平坦儀重現性與準確性之評估，本研究建立以實驗室法進行設備之認證流程，包含試驗進行程序之研擬及資料分析流程之建立，研究利用財團法人車輛測試研究中心之車輛耐久試驗設備，將慣性式平坦儀置於油壓缸伺服系統，輸入不同頻率與振幅之組合正弦波，以模擬不同路況等級之評估結果，並依循現有認證規範進行設備之評估；在資料分析流程建立部分，應用HHT於量測訊號之分析，可獨立選取設備量測主訊號，並降低車輛懸吊系統、市電及雜訊等對評估結果之影響。以實驗室法進行慣性式平坦儀之認證，可正確比對設備量測高程之誤差，有效獨立評估硬體之準確性，配合模擬路況進行分析軟體之認證，以評估分析軟體之正確性，可降低在人員操作誤差對慣性式平坦儀認證之影響。

Past studies have shown that most of the inertial profiler’s verification procedures are conducted through repeated profile measurements on test sections with known profiles. The evaluation takes account of errors from the profiler equipment (accelerometers, laser distance sensors, and displacement measuring sensors), analytical software (signal processing and IRI computations), and operator’s variables (tracking footprint and moving speed). In this study, a new method is presented to set aside the error factor of operator variables and concentrate only on errors of profiler instrumentation and analysis software. An experimental design was conducted at the Fatigue and Durability Test Laboratory of the Automotive Research and Testing Center (ARTC). Using a servo-hydraulic loading tester with indiscriminate sinusoidal vertical displacement frequencies and amplitude levels, the facility simulates several levels of roughness for different pavement test sections. This apparatus compiled sinusoidal vertical displacements of various frequencies and amplitude levels to generate the verification procedure. Besides, data analysis procedure is developed by using HHT and EMD, and it can reduce signal errors from vehicle suspension system and high frequency noise. Through the developed evaluation method, errors attributed to operator variables can be excluded, and errors induced by profiler and software can be further identified. This test procedure not only allows the comparison of calculated IRI values, but also provides a solid comparison of measured profile and the fed sinusoidal data point by point.