APPARATUS ABNORMALITY DIAGNOSIS METHOD AND APPARATUS ABNORMALITY DIAGNOSIS DEVICE

PROBLEM TO BE SOLVED: To obtain an abnormality diagnosis method with which it is possible to correctly detect a symptom of abnormality occurring in an object apparatus without being affected by a non-steady state change caused by spread of a state change in other apparatuses to the object apparatus, while taking into account the data collected from the object apparatus with regard to all state variables.SOLUTION: The present invention comprises the steps of: acquiring time-series data of a plurality of state variables of a plant having correlation with abnormality in an object apparatus; performing, for at least one state variable among the plurality of state variables, a preprocess to exclude, from the time-series data about the state variable, data of at least one state variable acquired within an exclusion period that is at least a portion of a transition state period in which the effect of a state change in other apparatuses constituting the plant spreads to the object apparatus, and obtaining abnormality diagnosing data about the plurality of state variables; and diagnosing abnormality in the object apparatus on the basis of the abnormality diagnosing data about the plurality of state variables.SELECTED DRAWING: Figure 3 【課題】全ての状態量について対象機器から収集されたデータを考慮しながら、他の機器の状態変化が対象機器に波及して起きる非定常的な状態変化の影響を受けずに、対象機器に異常が生じる兆候を正しく検知可能な異常診断方法を得る。【解決手段】対象機器の異常と相関がある前記プラントの複数の状態量の時系列データを取得するステップと、前記複数の状態量のうち少なくとも一つの状態量について、前記プラントを構成する他の機器の状態変化の影響が前記対象機器に及ぶ過渡状態期間の少なくとも一部である除外期間内に取得した前記少なくとも一つの状態量のデータを該状態量についての前記時系列データから除外する前処理を行って、前記複数の状態量について異常診断用データを得るステップと、前記複数の状態量の前記異常診断用データに基づいて、前記対象機器の異常診断を行うステップと、を備える。【選択図】図３