WASTE SUPPLY ABNORMALITY DETECTION METHOD, WASTE SUPPLY CONTROL METHOD, WASTE SUPPLY ABNORMALITY DETECTION DEVICE AND WASTE SUPPLY CONTROL DEVICE

To detect abnormality of waste supply to a combustion chamber at an early stage.SOLUTION: A waste supply abnormality detection device comprises: a learning part 45 allows learning of a prediction model using any one of multiple regression analysis, decision tree, random forest and deep learning using a steam generation amount in a predetermined period from a past time before a reference time until the reference time as an explanatory variable and using a prediction steam generation amount which is a steam generation amount at a future prediction time after the reference time as an objective variable, by using actual values of the explanatory variable and the objective variable; a prediction part 41a which inputs data on the explanatory variable to the learned prediction model by specifying a current time as reference time, and obtains a prediction steam generation amount; a difference calculation part 41b which calculates a difference in the steam generation amount by subtracting the prediction steam generation amount from the steam generation amount at a past time before the prediction time; and a detection part 41c which detects the presence/absence of waste supply abnormality on the basis of the difference.SELECTED DRAWING: Figure 2 【課題】燃焼室への廃棄物の供給異常を早期に検知する。【解決手段】廃棄物供給異常検知装置は、基準時刻より過去の時刻から該基準時刻までの所定期間の蒸気発生量を説明変数とし、前記基準時刻より未来の予測時刻の蒸気発生量である予測蒸気発生量を目的変数とする、重回帰分析、決定木、ランダムフォレストおよびディープラーニングのいずれかを用いた予測モデルを、前記説明変数と前記目的変数の実績値を用いて学習させる学習部４５と、学習させた予測モデルに現在時刻を基準時刻として前記説明変数のデータを入力し、予測蒸気発生量を求める予測部４１ａと、前記予測時刻より過去の時刻の蒸気発生量から前記予測蒸気発生量を減じて蒸気発生量の差分を算出する差分算出部４１ｂと、前記差分に基づいて廃棄物供給の異常の有無を検知する検知部４１ｃとを備える。【選択図】図２