貝塔衰減質量濃度監測儀之性能改善

Abstract

長期暴露於PM2.5會增加罹患呼吸系統疾病、心血管疾病和死亡的風險，為了人類健康的著想，使用自動連續監測設備監測粒狀污染物濃度是即時、有效率的方式。其中Beta gauge是環保署測站廣泛應用的儀器設備之一，主要由beta輻射源、輻射偵檢計數器和濾紙所組成，根據beta粒子通過微粒沉積的濾紙的衰減量來計算質量濃度，但是beta射源放射的beta粒子數目變異大，造成監測濃度的不穩定，導致beta gauge反應時間長且偵測下限高，其應用性受到限制，無法滿足未來PM即時監測與低濃度的需求。因此本研究將探討透過縮減採樣面積、增加本數與平滑方法等方法來改善beta gauge的反應時間與偵測下限，使其能有更快、更靈敏的數據反應短時間內的濃度變化，增加其應用性。

為降低採樣過程中外在環境的干擾，本研究使用氣膠產生系統來模擬環境中微粒的粒徑與質量濃度，測試beta gauge的性能並改善，透過擷取卡輸出每秒的beta強度轉換成每秒的質量濃度進行討論。針對以下三個部分來進行，第一部分是使用鋁板縮減採樣面積，在相同質量濃度下增加單位時間的累積厚度，當微粒的累積厚度增加可以增加beta強度的衰減量，使beta gauge在低質量濃度測量上更加靈敏。第二部分是在同一濃度的系統環境下用同一台beta gauge重複採樣，模擬多台beta gauge同時採樣之結果，增加beta強度的樣本數並取平均，降低beta強度的變異以獲得更穩定的即時濃度。第三部分是將即時濃度透過平滑方法來平滑每秒的質量濃度值，減少監測濃度因beta放射數目變異的跳動。最後以反應時間、監測濃度的標準差、訊噪比與偵測下限為評估指標，探討三個參數對beta gauge性能之影響。

性能測試結果顯示隨著使用的平滑時間愈短， beta gauge的反應時間愈短，但會造成監測資料的變異增加，此外也會導致偵測下限的增加。在固定採樣流率的情況下，增加樣本數可以大幅降低beta強度的變異，而縮減採樣面積可以增加beta衰減量，透過這兩個方法可以增加beta gauge監測資料的訊噪比，明顯降低使用低平滑時間所造成監測資料的高變異，達到更穩定的監測與更低的偵測下限。當使用的樣本數愈多，量測上可以更佳的穩定，但是當樣本數大於6組以上時，性能改善的效果已達平緩的趨勢，若持續增加樣本數會導致成本的浪費。

為達到更快、穩、靈敏的目標，建議使用4-6台連續式beta gauge同時採樣，來獲得更好的性能表現。當採樣面積縮減至0.4 cm2，樣本數增加至6組，以及使用30分鐘的平滑時間，可以將反應時間降低至22分鐘，標準差降低至2.67 μg/m3以及偵測下限降低至8.28 μg/m3。

Long-term exposure to PM2.5 increases the risk of respiratory diseases, cardiovascular diseases, and mortality. To protect human health, it is crucial to monitor particulate matter concentrations using automated continuous monitoring equipment. Among these, the Beta gauge is widely used for atmospheric PM monitoring. It consists of a beta radiation source, a radiation detection counter, and a filter. The mass concentration is calculated based on the attenuation of beta particles passing through the deposited filter. However, the number of beta particles emitted from the beta source varies significantly, causing unstable concentration measurements. This results in long response times and high detection limits for the Beta gauge, limiting its applicability for real-time PM monitoring and low-concentration detection. Therefore, this study aims to improve the Beta gauge's response time and detection limit by reducing the sampling area, increasing the number of samples, and applying smoothing methods. Faster and more sensitive data can reflect concentration changes in a short period, increasing its applicability

To minimize external environmental interference during experiment, this study used an aerosol generation system to simulate particle size and mass concentration in the environment. The performance of the Beta gauge was tested and improved by outputting the beta intensity per second and converting it into mass concentration per second. The study focuses on three parts: first, reduce the sampling area with an aluminum plate to increase the cumulative thickness per unit time at the same mass concentration. Increased particle accumulation thickness enhances beta intensity attenuation, making the Beta gauge more sensitive at low mass concentration. Second, the same Beta gauge repeatedly sampled at the same concentration in system, simulating multiple Beta gauges sampling simultaneously. It increased the number of beta intensity samples and averaged them to reduce beta intensity variation and achieve more stable real-time concentration measurements. Third, smoothing methods were applied to the real-time concentration to smooth out mass concentration data per second, reducing fluctuations in monitored data due to beta count variation. Finally, the Beta gauge's performance was evaluated in response time, standard deviation of monitored concentration, signal-to-noise ratio, and detection limit.

Performance results show that a shorter smoothing time reduces the Beta gauge's response time but increases data variation and detection limits. Increasing the number of samples significantly reduces beta intensity variation, and reducing the sampling area increases beta attenuation. These methods higher the signal-to-noise ratio of the Beta gauge's monitoring data, significantly reducing the high variation caused by a short smoothing time. Achieve more stable monitoring with lower detection limits. More samples lead to more stable measurements, but the improvement effect flattens when the number of samples exceeds six. If samples continue to increase, it results in wasted costs.

To achieve faster, more stable, and more sensitive targets, it is recommended to use 4-6 continuous Beta gauges for sampling at the same time. When the sampling area is reduced to 0.4 cm², the number of samples increased to six, and a smoothing time of 30 minutes is used, the response time can be reduced to 22 minutes, the standard deviation reduced to 2.67 μg/m³, and the detection limit reduced to 8.28 μg/m³.