機械故障診斷之非穩態信號特徵擷取技術

以往機械故障診斷環境在轉速固定或穩態狀態下，如核電廠主汽機所使用的泵、馬達等關鍵設備，應用頻譜分析方法（如傅立業轉換，Fast Fourier Transform，FFT）提取信號之頻譜特徵，作為辨識故障成因之根據。但對於瞬間轉速變化大的機械設備，如風力發電機齒輪箱或3C組裝用之機械手臂等，所產生的信號模式大部分為非穩態信號，其信號特徵會隨分析時間長度而平均化，無法被突顯，導致在故障診斷或辨識上之困難。為了改善此缺點，本文提出時頻階次譜方法。此方法結合短時傅立業轉換（Short-Time Fourier Transform，STFT）與轉速頻率階次方法，擷取非穩態狀態信號的階次特徵。此種信號特徵不因轉速變化而改變，可有效作為機械設備在非穩態狀態下之故障辨識。此外，本文將階次特徵輸入倒傳遞類神經網路（Back Propagation Neural Network，BPNN），進行齒輪-轉子實驗平台於非穩態狀態下故障診斷，驗證時頻階次譜方法之可行性。

關鍵字

非穩態信號；短時傅立業轉換；倒傳遞類神經網路；時頻階次譜方法

並列摘要

Previously, spectrum analysis (e.g., Fast Fourier Transform, FFT) was applied in failures diagnosis of fixed or steady state mechanism (e.g., pumps in nuclear power plant turbines, engines or other key equipment) to extract the frequency features as the basis in identifying the causes of failure types. However, mechanical equipment with instantly increasing speed variations (e.g., wind turbine transmissions or the mechanical arms used in 3C assemblies, etc.) mostly generates non-stationary signals, and their features must be averaged with analysis time, making it difficult to identify the causes of failures from the weakened signal properties. In this study, a time frequency order spectrum method combining with the Short-Time Fourier Transform (STFT) and speed frequency order method are proposed to capture the order features of non-stationary signals. These signal features do not change with speed, and are thus effective in identifying failures or faults in mechanical components under non-stationary conditions. Furthermore, Back Propagation Neural Networks (BPNN) and time frequency order spectrum methods are used to verify faults and superior diagnosis results in non-stationary signals of gear-rotor systems can then be obtained.

並列關鍵字

non-stationary signal ； short-time Fourier transform ； back propagation neural network ； time frequency order spectrum

國際替代計量

機械故障診斷之非穩態信號特徵擷取技術

全文下載

主題瀏覽