Data Mining And Knowledge Discovery In Database For Improved Industrial Safety

Abstract

Safety is crucial in oil and gas industries, it is normal practice to collect incident reports daily. The volume of such reports is always very huge, as a result, the interpretation poses a great challenge. Consequently, there is an urgent need for techniques that can intelligently assist in transforming such vast amount of data into useful information and knowledge. Data mining is part of the knowledge discovery process that offers ways of extracting non-trivial, previously unknown, and potentially useful information from data. It uses machine learning, statistical and artificial intelligent techniques to interpret the data and also discover hidden relationship and patterns in the data. In this research, the state-of-art of Data Mining and knowledge discovery, existing data mining tools and incident report management are reviewed A Data mining software application, VDMiner, for improved safety was developed for oil and gas industry. The work utilised the data mining standard process model by passing through business understanding, data preparation, modelling, evaluation, and development stages. The VDMiner uses the decision tree induction model on a dataset to analyze and predict incident scenarios that have the highest likelihood of occurrence at a particular location of the industry. The benefit of the application is that management in an organization can proactively deploy appropriate controls to prevent incidents/accidents to save life and reduce cost. VDMiner is a window based application and it is user-friendly. The installation, use and test results of the VDMiner are reported. Further areas of research on data mining are recommended.