Abstract
Plant projects, referred to as Engineering Procurement and Construction (EPC), generate massive amounts of data throughout their life cycle, from the planning stages to the operation and maintenance (OM) stages. Many EPC contractors struggle with their projects due to the complexity of the decision-making processes, owing to the vast amount of project data generated during each project stage. In line with the fourth industrial revolution, the demand for engineering project management solutions to apply artificial intelligence (AI) in big data technology is increasing. The purpose of this study was to predict the risk of contractor and support decision-making at each project stage using machine-learning (ML) technology based on data generated in the bidding, engineering, construction, and OM stages of EPC projects. As a result of this study, the Engineering Machine-learning Automation Platform (EMAP), a cloud-based integrated analysis tool applied with big data and AI/ML technology, was developed. EMAP is an intelligent decision support system that consists of five modules: Invitation to Bid (ITB) Analysis, Design Cost Estimation, Design Error Checking, Change Order Forecasting, and Equipment Predictive Maintenance, using advanced AI/ML algorithms. In addition, each module was validated through case studies to assure the performance and accuracy of the module. This study contributes to the strengthening of the risk response for each stage of the EPC project, especially preventing errors by the project managers, and improving their work accuracy. Project risk management using AI/ML breaks away from the existing risk management practices centered on statistical analysis, and further expands the research scalability of related works.
Highlights
Engineering, Procurement and Construction (EPC) is a form of contract in which a prime contractor guarantees the obligations of engineering, the procurement of materials and equipment, construction, and a warranty for a plant project, as a lump-sum turnkey base in most cases [1]
The range of design cost estimation to be predicted this in the past were collected from EPC contractors for onshore and offshoreinplants, and study is limited to the number of design man hours (M-H) required from the bidding to model for the estimation of the design M-H was developed using this information
The ML model applied to the design package module and predictive maintenance module of this study was tested through the Engineering Machine-learning Automation Platform (EMAP) system, and due to the space limitation of this paper, the application results of all of the models appearing on the system are presented in a table in the validation for each module
Summary
Engineering, Procurement and Construction (EPC) is a form of contract in which a prime contractor guarantees the obligations of engineering, the procurement of materials and equipment, construction, and a warranty for a plant project, as a lump-sum turnkey base in most cases [1]. It is necessary to study a system that analyzes risk factors by applying AI and big data technology to support the utilization of the engineering data generated in each stage of the EPC project and decision-making through it. ML platform, the Engineering Machine-learning Automation Platform (EMAP), that applies various ML models and AI algorithms for risk analysis at each stage of the bidding, design, construction, and OM. These five modules support the risk analysis for plant projects at each stage of bidding, design, construction, and OM. The project data collected for this study were used for the training of the corresponding module of EMAP with embedded machine learning model Based on this training, it was designed to generate a final prediction when the user inputs a new document
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have