引導式測驗 · 中等 等級

預測性維護測驗中的人工智慧

了解機器學習方法如何在實踐中使用狀態監測資料來檢測異常並估計設備壽命。

問題 1 的 8

與預防性維護相比,預測性維護主要解決什麼問題?

預覽每個問題8 問題

當您參加測驗時,就會出現答案和解釋。

  1. 與預防性維護相比,預測性維護主要解決什麼問題?
    • 僅在需要時進行修復,從而避免以固定時間表更換健康零件
    • 它消除了對任何感測器的需要
    • 它使機器運行得更快
    • 它完全消除了對技術人員的需求
  2. 「剩餘使用壽命」(RUL) 的估計值是多少?
    • 一台機器的總成本
    • 組件失效前剩下多久
    • 馬達每小時消耗的能量
    • 設備上的感測器數量
  3. Which signal is collected in the Case Western bearing fault dataset for rotating-machine diagnostics?
    • Patient electrocardiograms
    • Satellite color channels
    • Vibration acceleration
    • Text sentiment scores
  4. 自動編碼器通常如何支援 PdM 中的異常檢測?
    • 它會自動按名稱標記每種故障類型
    • 它產生新機器
    • 它取代了振動感測器
    • 它僅針對健康數據進行訓練,並將高重建錯誤標記為異常
  5. What edge-computing approach does the cited condition-monitoring paper study?
    • Federated autoencoder training using vibration data at devices near the monitored machines
    • Centralizing all raw data in a public image database
    • Training a language model on maintenance manuals only
    • Replacing sensors with a periodic questionnaire
  6. How does NIST define a digital twin?
    • A maintenance calendar with no model of equipment
    • A virtual representation of a physical or perceived real-world entity
    • A physical duplicate machine kept in storage
    • A guarantee that a machine will not fail
  7. What is the NASA C-MAPSS dataset used to evaluate?
    • Cancer grade classification from biopsy slides
    • Speech recognition in noisy rooms
    • Turbofan engine degradation and remaining-useful-life prediction
    • Global climate averages without time series
  8. How should an anomaly alert from a predictive-maintenance model be interpreted?
    • As a guaranteed failure date
    • As an automatic repair prescription
    • As evidence that no human review is needed
    • As a signal for inspection, not proof of a specific fault or failure time