Identification of Abnormal Monitoring Data of Unmanned Ship Diesel Engine Based on Information Entropy

The harsh working environment of diesel engine and its condition monitoring are easily disturbed by noise, so that the collected diesel engine operation data are often accompanied by abnormal values. For unmanned ships, the existence of these abnormal values has a serious impact on the evaluation of diesel engine health status and the accuracy of system independent decision-making. To solve this problem, this paper presents a method of identifying abnormal monitoring data of unmanned ship diesel engine based on information entropy. Firstly, the consistency of diesel engine monitoring data is described by calculating the information entropy of monitoring data. On this basis, the anomaly degree of each monitoring data is calculated and compared with the anomaly threshold to complete the confirmation of abnormal data. In this paper, the abnormal output data of the sensor is obtained based on a certain diesel engine test-bed, and the feasibility of mining diesel engine abnormal monitoring data based on information entropy is verified.