Adaptive Learning for Zero Defects in Building Construction

Abstract

"According to the industry-led, Get It Right Initiative, the total cost of avoidable errors in the UK construction industry is £10-£25Bn per annum (or 10%-25% of project cost).

If these errors could be reduced, then buildings could be built at lower cost and more buildings could be constructed with the same labour and materials, thus increasing productivity.

Sometimes defects (caused by errors) are not detected when the building is completed and handed over to the occupants, and they persist for many months or even years. So by reducing these defects, the quality of the environment in which people live and work will be improved. An undiscovered defect may also endanger the occupants of a building.

Buildings may also use more energy and water than they were designed to because of defects, leading to higher bills, energy waste and unnecessary carbon dioxide emissions to the atmosphere.

Defects often occur due to insufficient site supervision and poor communication with site operatives. Sometimes operatives are reluctant to ask for clarification because they are concerned about appearing ignorant or inadequate. This can result in tasks being performed incorrectly leading to defects.

Defects (sometimes called 'snags') are often discovered during an inspection. When a defect is discovered it can be a time-consuming and costly process to correct the problem.

The aim of this project it to achieve the construction of buildings with zero defects. The zero defects philosophy was promoted in the automotive manufacturing sector more than 30 years ago and we now have better quality cars, that cost less (relatively) than those made in the 1970s.

Our project aims to achieve fewer defects by preventing their occurrence, rather than by detecting and fixing them. This is far preferable because it will lower the overall cost of construction and also prevent delays.

Our proposed innovation (to be based upon this feasibility study) will use computer technology (mobile and cloud) and Artificial Intelligence (AI) to check that a site operative understands the task that they are being asked to perform. If they don't, it will attempt to explain it to them in a way that they can understand. If the system is not satisfied that the instructions have been understood it will alert a supervisor.

In this way it will prevent defects that might have occurred because of a lack of understanding. As the system is used it will learn more about defects and defect prevention."

Lead Participant

Project Cost

Grant Offer

Tr Control Solutions Limited, Surrey £184,780 £ 129,346
 

Participant

Anglia Ruskin University, United Kingdom £63,301 £ 63,301

Publications

10 25 50