Remote monitoring and failure prediction of guiding elements and diverting pulleys in passenger elevators

Thumbnail Image
Journal Title
Journal ISSN
Volume Title
Insinööritieteiden korkeakoulu | Master's thesis
Product Development
Degree programme
Master's Programme in Mechanical Engineering (MEC)
46 + 0
Accelerated urbanization has lead to the rising height of buildings and demand for intensive high performance of elevators in recent years. Consequently, condition monitoring has become a highly desirable capability as the complexity of elevator systems increased. The goal of this study is to develop a monitoring method for elevator components which are subjected to mechanical degradation and failures. The method is capable of indicating the current health condition, predicting future failure as well as detecting emerging issues during operation. Studies of the fundamental principle of elements of condition monitoring such as measurement and measuring equipment, remaining useful life models laid the foundation for new method developing. Moreover, there were reviews of the implementation of health management systems in aerospace and marine industry. A prototype was built from the inductive sensor and open sources embedded system. The device has been installed in two different elevators for data acquisition. Basic data visualization and analysis models were employed for current health state assessment and failure trend prediction. The results include validation of the condition monitoring method and prediction of time-to-failure. Arithmetic means of displacement data determined operating condition whereas the linear regression model was used to predict failure event. Moreover, while suggesting the potential usefulness of the method for system condition assessment, the analysis of the data also exposed challenges inconsistency of the measuring method, data filtering technique as well as large data size requirement.
Kuosmanen, Petri
Thesis advisor
Sorvali, Ville
condition monitoring, elevator, displacement measurement, data analysis, regression, RUL
Other note