基于具有明显测量误差的辐射图象的二维炉膛温度分布重建和快速特征识别研究

STUDY ON RECONSTRUCTION AND FAST CHARACTERISTIC RECOGNITION OF TEMERATURE DISTRIBUTIONS IN A 2-D FURNACE VIA RADIANT ENERGY IMAGES WITH SENSITIVE MEASUREMENT ERRORS

采用辐射图象处理进行二维炉膛温度分布重建是一个典型的逆问题。由一个数字代摄像装置 (CCD)镜头获取的信息重建温度场会使重建结果对误差过于敏感。本文采用4个CCD镜头来获取更多有效信息 ,将温度分布重建问题转化为一种线性规划求解问题。通过给定温度的范围来加强对问题的约束 ,并采用一种内点法算法建立了一种二维炉基金项目 :国家自然科学基金资助项目 ( 5960 60 0 7)膛温度分布重建和快速特征识别方法。对单峰和双峰温度分布的计算结果表明 ,此方法可以显著提高整个系统在辐射图象具有明显测量误差情况下的重建精度 ,满足实际应用的要求。该方法很容易推广到电站锅炉炉膛三维温度场重建及其它领域的应用中。

Reconstruction of temperature distribution through radiant image processing in 2 D furnaces is a typical inverse problem, and it is very sensitive to measurement errors by only one CCD camera. Radiant images captured by four cameras are used to add more useful information, and the inverse problem is converted into a kind of linear programming problem. Through adding constraints of the temperature bounds in its distribution, an interior point algorithm is adopted to solve the linear programming and forms a method of reconstruction and fast characteristic recognition of temperature distribution in 2 D furnaces. Calculation results for single peaked and double peaked temperature distributions in a 10m×10m furnace indicate that this algorithm can remarkably improve the accuracy of the reconstruction results with sensitive measurement errors, and can satisfy completely the requirements for application. This method can be easily extended into reconstruction of 3 D temperature distributions of utility boiler furnaces and applications in other related fields.