To Örebro University

Automating the detection of processing conditions that may lead to defects in the wire during the wire drawing process is of high interest to the industry. Current practise is based primarily on operator experience. Increasing demands on product quality and process robustness emphasises the need for development of robust in-process detection methods. This work is focusing on investigating the potential of using vibration monitoring to detect process deficiencies or variations that may lead to defects in the product. Wire drawing of a carbon steel in different lubricating situations was used to investigate vibration signal response together with force measurements and surface investigation of the wire product. The results show that vibration measurement is capable of detecting loss of lubrication that leads to poor surface quality of the wire.

Wire drawing is a cold work metal forming process which is dependant of a functional lubrication process. If the lubrication fails, there is a risk that both the tools and the produced wire will be damaged. Process monitoring of wire drawing is rare in today’s industry since there are no commercialised methods that deliver consistent results. In this paper, a method for monitoring of the wire drawing process is proposed and evaluated. A thermal imaging camera was used for acquiring thermal images of the wire as it leaves the drawing tool. It was found that the proposed method could capture changes in the wire drawing process and had correlation to the drawing force. An equation for estimating the friction condition between the wire and the drawing die using the wire temperature was also proposed and evaluated against experiments. The results showed that the new equation produced results that correlated well to results obtained using a conventional equation that use drawing force.

Wire drawing is a cold metal forming process where a wire is drawn through a series of dies, reducing the wire dimension, and enhancing the material properties. The wire drawing process requires lubrication to function, as discrepancies in the lubrication can cause failure of the entire process. Such incidents can be costly and there is a need to monitor the process so that changes in the lubrication can be detected, and addressed, before failures occur. The aim of this work was to determine whether a CCD-camera could be used to monitor the wire drawing process. The purpose of the monitoring was to detect if the process was about to fail. In this work the failure of the process was initiated by removal of the lubricant, causing galling between the wire and the die. The signal from the CCD-camera was compared to the signal from a drawing force measurement that clearly indicates when the friction in the die increases, which in turn indicates imminent failure of the process. It was found that the CCD-camera signal clearly indicated the removal of lubricant, and thus failure. In this work, the CCD-camera was tested on two different wire materials and two different lubricant, both with positive results. All tests were performed in an industrial wire drawing setup.

In-situ evaluation (monitoring) of the wire drawing process is often performed manually in today’s industry which is a difficult task for the operator, requiring both time and experience. Previous research at Örebro university has been performed to identify and evaluate automated in-situ monitoring methods for the wire drawing process. From this research, several methods that can be applied for monitoring purposes have been identified. However, the advantage of using multiple sensors has not yet been investigated. How data from different monitoring sensor signals correlates with each other and if they can be combined to obtain a better understanding of the wire drawing process will be investigated and discussed in this work. Four different sensor signals; vibration, wire temperature, brightness of the wire surface and drawing force, will be compared and evaluated in wire drawing experiments where the process conditions are controlled. The results show that all the evaluated sensors indicate similar to deviations in the lubrication process, however, some problems could only be detected by some sensors. Using multiple sensors can have advantages in both detection and misrepresentation of problems and might be used to classify specific types of problems in the process.