On-board Feature Extraction for Clutch Slippage Deviation Detection
| Autor: | Källström, Elisabeth |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2015 |
| Předmět: | |
| Druh dokumentu: | Text |
| Popis: | Construction equipment companies continuously upgrade their products to meetcustomer demands, staying competitive with market challenges as well as improving sales and profits. With increased complexities in heavy duty machines today, up-time is considered an important aspect of the construction equipment business because it reduces warranty and service cost, while increasing sales and overall customer satisfaction. Therefore, a substantial amount of research is directed towards the development of intelligent machines which are capable of automatically monitoring the health of different components in the machine.Heavy duty construction equipment is often equipped with automatic transmissions, with multiple disc wet clutches, transferring torque from the engine to the gearbox enabling automatic gear ratio changes. The wet clutches in Volvo Construction Equipment vehicles may be considered as a crucial component of the driveline and failure may result in costly downtime. The frictional characteristics of wet clutches are crucial for the ultimate performance because they define clutch slip time during engagement. Furthermore, a wet clutch is considered to have failed when it can no longer transmit the desired torque. The level of torque transfer in wet clutches is controlled by the generated friction. Hence, clutch slippage is the result of diminishing frictional characteristics of wet clutches. Accurately monitoring slippage in wet clutches provides an indication of the health of the clutch material.However, many of the factors that influence the frictional characteristics of the wet clutch are only possible to measure in a test rig and not in actual machines.In this thesis the gap in condition monitoring of automatic transmission clutchesin an actual machine is addressed. A methodology for monitoring the health of the clutch material on-board the machine using the available CAN-bus signals via an on-board Data Stream Management System (DSMS) has been developed. The feature extraction methods utilized in the condition monitoring are based on Moving Average Square Value Filtering (MASVF) and a measure of the fourth order statistical properties of the CAN-bus signals implemented as continuous queries over data streams. Results show that the feature extraction methods provide an indication of clutch slippage deviations.This thesis also includes an investigation of clutch slippage detection from driveline vibrations based on spectrogram and spectral Kurtosis methods. Fastelaboratoriet - VINNEXC |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|