Train performance prediction method based on measurement space assisted robust scheme

Abstract

The invention relates to a train performance prediction method, in particular to a train performance prediction method based on a measurement space assisted robust scheme. The objective of the invention is to solve the problem of low accuracy of train performance prediction. The method comprises the following steps: 1, offline stage preprocessing and data modeling: collecting and preprocessing train migration system performance data to obtain preprocessed train migration system performance data; establishing a train non-interference migration system data model based on the preprocessed train migration system performance data; establishing a soft measurement data model based on the preprocessed train migration system performance data; 2, off-line stage parameter identification: identifying a train non-interference migration system; identification of soft measurement is carried out; based on identification of a train interference-free migration system and identification of soft measurement, identification when interference exists is carried out; and 3, an on-line stage: collecting performance data of the train migration system for performance prediction. The method is applied to the field of train performance prediction.

Images