Application of Natural Language Processing (NLP) and Text-Mining of Big-Data to Engineering-Procurement-Construction (EPC) Bid and Contract Documents

• Published in: 2020 6th Conference on Data Science and Machine Learning Applications (CDMA) pp. 123 - 128
• Main Authors: Kim, Youyi, Lee, Junghyun, Lee, Eul-Bum, Lee, Ju-Hoon
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.03.2020
• Subjects: Companies; Contracts; Decision making; Decision-making support system; EPC; ITB Analysis; Maintenance engineering; Natural Language Processing(NLP); Risk management; Semantics; Text analysis
• DOI: 10.1109/CDMA47397.2020.00027

Influenced by the innovative technological advances made in recent years, a number of major economies are increasing their national investment in building decision support systems applicable to the industrial sector. This study was conducted with government support to analyze big data in engineering accumulated from the past, develop tools for forecasting and risk management, and use them to build systems that support decision-making. The target of system development for this study is the industrial engineering sector, also known as Engineering- Procurement-Construction (EPC), and is largely divided into five modules, as follows. First, engineering design cost prediction targeted accurately predicting Man-Hour (M/H) and provided useful references to contractors and project managers by developing models that predict design costs in the engineering field. Using the M/H prediction model developed by this research, M/H under similar environments or conditions can be predicted and solutions prepared on the project start stage to manage potential risks. Second, engineering design error analysis was conducted with the purpose of helping project implementation risk management by developing a system to analyze actual design error data that occurred in past projects in order to minimize and prevent design errors. Third, engineering design change analysis was conducted to analyze amounts that could be added by changing order and preparing countermeasures in advance due to design changes occurring during the project. The system was built to support decision-making as a result of utilizing accumulated data. Fourth, engineering Invitation to Bid (ITB) analysis established a support system to efficiently perform ITB analysis and carry out new processes that automatically detect and present provisions that could pose risks. Natural Language Processing (NLP) techniques were applied to conduct the analysis, and Subject-Matter Experts (SMEs) verified the quality during the process of establishing the procedure. Fifth, equipment maintenance cycle forecasting was conducted to predict the demand for repairing items and ensure that equipment requiring maintenance could be prepared before the predicted malfunction. This paper first introduces the five modules of decision support systems in a wide range, then details the results of the fourth module, the Engineering ITB Analysis step by step.