本研究針對燃燒爐煙氣氧含量控制系統進行效能評估。控制煙氣的氧含量保持在設定點,使燃燒爐的燃燒效率保持在最佳狀態,減少溫室氣體的排放,廢氣處理成本得以降低。然而燃燒爐系統受到未知的外界干擾影響,使煙氣氧含量產生變異,因此評估目前的控制系統以決定是否修正控制器,使燃燒爐系統的保持在預期的效能。為了要評估目前燃燒爐系統的控制效能,我們應知道目前燃燒爐的最小變異量。本研究首先針對氧含量的單變數回饋控制系統建立燃燒爐系統的最小變異量,此為系統的變異量極限值。但最小變異量控制受限於實際的控制器,無法實際應用在真實的燃燒爐系統,所以本研究將推導實際控制器可達到的最小變異量,同時提出以SOS規劃(Sum of squares programming)求解此可達到最小變異量。 在燃燒爐系統中,由於規模越龐大,受到質傳與熱傳的物理限制,造成氧含量量測的時間延遲增加。若燃料進料與空氣進料組成發生變異,因量測時間延遲使得最初的干擾控制器不能及時修正,直到較大的干擾產生。由於火焰的行為能及時了解目前燃燒的狀況,本研究提出利用火焰影像控制,由於火焰影像的資訊太大,使用MIA (multivariate image analysis)從火焰影像萃取特徵,而後以此特徵以進行火焰影像控制與其控制效能評估。 雖然在火焰影像控制中,已能及時消除影響火焰影像的外界干擾,但火焰影像的變動頻率很大,擷取的影像無法反應干擾的資訊。即使以高速取樣的火焰影像,操作的控制閥並不能完全接受此快速的變動。為克服此問題,本研究將提出串級式控制,將火焰影像控制做為內環路,外環路為氧含量控制,建立火焰影像與氧含量的串級控制設計及其最小變異量。最後藉由實驗測試比較上述三種控制系統的優劣。
This research studies the performance assessment of oxygen content control of flue gas in combustion. Maintaining the oxygen content in flue gas at the set point enables efficient combustion, reducing greenhouse gas emissions and waster gas treatment costs. However, the combustion system is affected by unknown disturbance which may alter the oxygen content in flue gas. Thus, the control system should be assessed to decide whether the controller should be amended to maintain the expected performance of combustion. In order to assess the current performance of the combustion control system, the minimum variance of the combustion system is required. In this study, the minimum variance for a single variable feedback control system for oxygen content control is determined. It is set as the limit of variance. Due to the limitation of the controller, minimum variance control cannot be directly applied to a real combustion system. Achievable minimum variance control is used instead and is obtained by the use of sum of squares programming. In the combustion system, due to the large scale of the system, the physical limitations resulting from mass transfer and heat transfer cause delay in the measure of the oxygen content of the system. The delay results in the controller unable to deal with change in the composition of the fuel and air until greater disturbance occur. The flame characteristic, on the other hand, allows immediate detection of the current combustion status. In this study the real time flame status is used for oxygen content control in flue gas. The huge amount of data from the flame image is handled by multivariate image analysis (MIA) for the flame image feature extraction to be used in image control and control performance assessment. Although flame image control is capable of promptly eliminating disturbance of the flame, the rapid changes in flame indicate that the captured image cannot reflect the disturbance. Even with high speed capture of the flame, the valve will handle such rapid changes. To this end, a cascade control method is proposed. Its inner loop is image control, and the outer loop, oxygen content control. The minimum variance of the cascade control system is evaluated. An experimental study is used to compare the advantages and disadvantages of the above three control systems.
為了持續優化網站功能與使用者體驗,本網站將Cookies分析技術用於網站營運、分析和個人化服務之目的。
若您繼續瀏覽本網站,即表示您同意本網站使用Cookies。