- 學位論文
車輛引擎故障診斷整合振動擷能技術開發
Technique Development of Vehicle Engine Fault Diagnosis Integrated with Vibration Energy Harvest
摘要
近年來為符合相關法條規範已有許多車輛引擎診斷系統陸續被開發，例如歐洲車載診斷系統(European on-board diagnostics, EOBD) 或強化型第二代車載診斷系統(on-board diagnostics II, OBDII)。本論文嘗試以加速規(accelerometer)研究車輛引擎的振動信號擷取及解析其特徵訊號，並應用訊號處理的技術於實車引擎故障分析與探討。佐證階次追蹤技術(order tracking technique)能有效建構車輛引擎振動訊號之初步動態特徵，引入快速傅立葉轉換(fast Fourier transform)及小波轉換(wavelet transform)理論探討在時–頻域分析(time–frequency domain analysis)的振動態樣；接續應用相類似的觀念，針對引擎進氣歧管壓力變異作為引擎運轉狀況的偵測訊號，結合離散小波轉換(discrete wavelet transform, DWT)並應用人工類神經網路(artificial neural network)技術建構故障診斷的專家辨識系統(expert detection system)。而車輛運動時會提供一固定週期之衝擊振動能，而此現象則是本論文探究的另一重點；文中闡述如何應用壓電智慧型材料研製換能器，並建構一槓桿機構模擬實車運動時的振動實驗平台，可將車輛運動時的週期性衝擊振動能加以發電回收，產生無汙染的振動能量擷取(vibration energy harvesting)系統，實現再生綠能發電的研究成果與未來應用方向。文中亦說明如何引用有限元素分析軟體ANSYS來模擬輔助開發，使製作出之壓電換能器能事先完成機械性質及電性上的模擬，並實驗分析此智慧型壓電換能器在操作的振動頻率範圍內相關發電效能之特性。
並列摘要
In recent years, several On-Board Diagnostics of automotive engines, such as European On-Board Diagnostics (EOBD) and On-Board Diagnostics II (OBDII stands for 'Enhanced On-Board Diagnostics, Second Generation'), have been developed to the fault diagnostics of vehicle engines because of legislative regulations. To study the dynamic characteristics of engines, the vibration patterns were obtained through signal acquisition with accelerometers and analyzed via data processing. Also, the several techniques of signal process would be adopted to investigate engine faults. After the order tracking technique could construct the original dynamic patterns for engine vibrations; therefore, the fast Fourier transform and the wavelet transform were introduced to extract the vibrating feature based on the time–frequency domain analysis. Applying the similar concept, the pressure variations on an intake manifold of an internal combustion engine should also detected as fault diagnosis. These pressure signals were decomposed and reconstructed by the discrete wavelet transform (DWT) and the engine malfunction could be recognized via various techniques of artificial neural networks. Thus, the expert detection system is developed for engine fault detection. The phenomena of a periodical impact oscillation are the energy resource through the vibration harvest with piezoelectric smart materials. The power generation is available and experimental analysis is another design point of finite element ANSYS method in this paper. The testing platform was built up and used to prove the conversion efficiency from vehicle vibration energy to electrical power output by using a lever mechanism to simulate the oscillating situation from a vehicle operation actually, called as a vibration energy harvesting system.