化工生產過程中，開展測量數據的誤差偵破與校正方法的研究具有十分重要的意義。針對焦化碳一過程中測量變量的穩態數據校正，采用殘差檢驗方法進行數據協調和過失誤差的識別、偵破，通過兩層次變換進行數據分類，從而消除了已有算法中出現奇
異矩陣的情況。校正后的變量滿足物料和元素平衡要求，誤差偵破的結果為儀表故障的排除等提供了參考依據;數據校正則提高了生產效率。
關鍵詞:兩層次 投影矩陣 數據校 正過 失誤差 偵破
Abstract: In process of chemical production, it is significant to develop error detection of measured data and research on calibration method.
Aiming at stable state data correction in process of 1-carbon,場using residual detection method, the data was coordinated, the unpremeditated
errors were recognized and detected; the data sorting was implemented through two-level transformation method; thus the unusual matrix in previous
algorithms was eliminated. After correction, the variables met the demands of balance between materials and elements. According to the
results of error correction, the reference is provided for fixing the faults of instruments, and the productive efficiency is enhanced場data correction.
Keywords:T wol evelp rojectivem atrix Datac orrection Unpremeditatede rrord etection