Application of Artificial Intelligence to Automate Construction Materials Data Classification

Abstract

The applications of Big Data analytics and Artificial Intelligence (AI) have gained a widespread attention in the construction industry in recent years following the promulgation of Industry 4.0. In the realm of construction research, AI has been utilised widely in areas such as structural design optimization, resource and equipment planning, and project scheduling. The research presented in this paper is aimed to utilise AI to assist with the automatic classification of the large volume of construction material orders created by users through an online marketplace website. Such big data of material orders contained numerous errors (e.g. typographical errors and incorrect units) that were extremely time consuming to correct before the datasets can be used to for further business intelligence analysis. In this research, the dataset was obtained from a business-to-business e-commerce company in Thailand, namely BUILK. The data from BUILK was the construction materials purchase orders created by BUILK's customers through its website, which contained hundreds of thousand unorganized records. In this study, Artificial Neural Networks (ANNs) was applied to automate the categorization of approximately 220,000 records of reinforcement steels orders. The ANNs model was developed and trained using over 32,000 records, with approximately 92 percent of prediction accuracy. The model automatically categorized the steel reinforcement data into 11 groups; Deformed Bars, Round Bars, Wire mesh (Deformed Bars), Wire mesh (Round Bars), Stirrup, Anchor, Material and Others. The outcome of this research helped the company to easily analyze the data to generate insights for its business management and development.

Opportunities and Adoption Challenges of AI in the Construction Industry: A PRISMA Review

“…Data accessibility: The demand for highly personalized projects continues to rise, and the accessibility of data in real-time has become paramount. Stakeholders can solve problems more efficiently by communicating with multiple project teams and meeting changing customer anticipations using real-time data [86]. Furthermore, the growth of IoT increases the ability to record data passively through portable devices [52,86].…”

“…Stakeholders can solve problems more efficiently by communicating with multiple project teams and meeting changing customer anticipations using real-time data [86]. Furthermore, the growth of IoT increases the ability to record data passively through portable devices [52,86]. This allows project managers to automatically prioritize issues and rate subcontractors according to a risk score so that construction managers can work closely with high-risk teams to mitigate risks [40].…”

Reality Capture in Construction Project Management: A Review of Opportunities and Challenges

“…Therefore, the provision of real-time digital data being made available remotely for decision making is imperative. RC has the capability to provide ongoing project data [3,65,66]. However, the request for extremely tailored projects keeps growing, and real-time data access has become critical.…”

“…The presented visual capabilities and opportunities of RC technologies allow for increased communication among different parties participating in the building process when observing and giving comments and judgements for a given project phase. RC is among the successful means of gathering site data [2,17,65,91].…”