Online boiler combustion optimization method based on historical working conditions

The invention relates to a boiler combustion online optimization method based on historical working conditions. The boiler combustion online optimization method comprises the following steps: defining optimization targets as boiler efficiency and NOx emission concentration; operating data of the boiler are collected, abnormal values are removed, and steady-state working conditions are judged; then, a K-means clustering method is adopted to represent the load of the external constraint condition of the boiler and the relative coal quality coefficient as division indexes, and the boiler operation working conditions are divided; and finally, optimizing the boiler efficiency and the NOx emission concentration by adopting a multi-target fuzzy optimization method. According to the method, when the K value of the K-means clustering is determined, the elbow amplification is adopted to determine the K value, and the data range participating in classification is pre-judged through data distribution, so that the accurac