摘要 呼吸器在重症醫學領域中具有重要的功能，其最大的功用是給予呼吸衰竭的病人機械通氣，提供支持生命，呼吸器除了提供機械通氣之外，具備有警報功能，當病人使用呼吸器時有偵測到異常的呼吸危象，會產生若干種警報，提醒臨床醫療人員，病人處於異常狀態，其中又以高壓警報最為常見，其常見的原因為病人呼吸道有痰液阻塞亦或咳嗽。痰液量的變化，會跟肺部發炎和呼吸器脫離成功與否有關。但是目前仍沒有發展一個系統工具可以估算呼吸器病人痰液量，以進一步提供臨床治療病人的參考。 本論文目的為量化呼吸器「流速-時間」波型上之震盪值並結合痰液濃度來做為智慧型呼吸器病人痰液之估測設計，並應用模糊理論(Fuzzy Logic)為基礎，結合大量數據組成歸屬函數與規則庫，使用最大-最小推論引擎，且利用最大值平均去模糊法(MOM)，估測出呼吸器病人痰液量，輸出至LabVIEW人機介面，將此參數推廣至作為病人脫離呼吸器的重要判斷依據與床邊監視喘的指標之一，並將病人呼吸道中痰液抽除乾淨，亦可減輕醫護人員的負擔。 本論文的主要設計是估測出呼吸器病人痰液量。利用模擬肺，及使用模擬痰液，計算在不同濃度痰液之呼吸流速震盪量，以進一步透過設計的智慧型呼吸器病人痰液之估測系統，來呈現痰液估算量。並結合呼吸監視系統，亦可將痰液估算量寫入電子病歷中，這樣必會大幅提升醫療人員的工作效率。 ;Abstract Mechanical ventilator performs essential functions in critical care medicine. Mechanical ventilator provides respiratory failure patients with mechanical ventilation which supports patients to maintain gas exchange. The thesis designed a system to estimate sputum volume which is needed in chest intensive care unit. The Sputum Volume Estimate System will be initiated immediately when the ventilator high pressure alert go off. The proposed sputum volume estimate system is based on the vibration (VIB) quantified from the breathing flow rate line. The fuzzy logic algorithm is applied to estimating the sputum volume. The inputs of the fuzzy logic include “Concentration” and “VIB value”. The fuzzy logic algorithm is divided into three steps to calculate the sputum volume. First, the concentration and VIB value are fuzzified. Then the inference engine applies the Min-Min-Max inference engine with the rule base to create twelve fuzzy rules. Last, the Volume value fuzzification utilized the Middle of Maximum Defuzzification method. The fuzzy logic algorithm outputs the sputum volume value interval. Finally, the sputum value and the indicator are shown on LabVIEW human machine interface. This thesis primarily designs the process of measuring sputum volume automatically. The proposed sputum volume estimate system could reduce unnecessary sputum suction. Then the system can keep the patient from suffering the side effects is caused by unnecessary sputum suction.
