Hits:

Date of Publication:2022-10-10

Journal:自动化学报

Issue:3

Page Number:462-469

ISSN No.:0254-4156

Abstract:The paper presents a new detection method of abrupt change based on extended sequential probability ratio test (ESPRT) innovation process. This method can convert a parameter model into a non-parameter model using innovation theory. So, the method can be widely applied to practice in combination with ESPRT. The paper uses nonlinear temporal series based on BP neural network to establish the innovation model and leakage detection model of the pipe. It is proved that the detection method can effectively online detect the leak, using the collected data in the laboratory.

Note:新增回溯数据