Analysing qualitative data systematically using thematic analysis for deodoriser troubleshooting in palm oil refining

Abstract

Troubleshooting the deodoriser in palm oil refining is challenging because of complex design of vacuum system. Not all the required information to develop a good troubleshooting system can be obtained from documents and the literature. A mistake in corrective action may cause failure in plant operation and final oil quality. Since experienced plant workers have in-depth knowledge on the specific equipment and process that they are in-charge of, systematic knowledge acquisition from them can be used to develop a fuzzy logicbased troubleshooting system. This system can help other less experienced staff in identifying root cause of deodoriser failure before taking any corrective action. The qualitative data obtained from interviewing the plant experts must be analysed systematically to ensure that the information obtained are trustworthy. Therefore, this study implemented thematic analysis to sort knowledge based on a theme of the root cause of deodoriser failure. The thematic analysis consists of 5 steps, which are data familiarisation, code generation, theme searching, reviewing and description. Fifteen plant workers (includes engineer, supervisor, technician and operators) were interviewed to develop themes. The useful information is captured from the interview session through code generation. Two senior managers verified the final themes, and the theme is further described based on the theoretical knowledge and process description. The final themes presented as thematic tree provide fault propagation, while troubleshooting table has information on root cause and its necessary action. The main themes presented the root cause of deodoriser failure based on critical location, i.e. ejector. The documented thematic analysis can assist inexperienced workers to learn the prior knowledge of deodoriser failure sources from troubleshooting table and thematic tree.