Investigate safety and quality performance at construction site using artificial neural network

Abstract

Quality inference of a construction project agenda might be an exigent task for project stakeholders. Construction superintendent is decisive to eventual site security performance. In the United States, the OSHA 30-hour training is becoming the de facto standard for supervisor security competency. We concentrate on gap by recognizing the essential knowledge-based security competencies that are most important for the front-line construction supervisor and precedence them for the first time. The intention of the work is to frame an Artificial Neural Network (ANN) with the assist of the optimization techniques. The ANN is utilized to predict the number of rework Work-hrs per $1M in Scope, a number of rework workers (works)-hrs per 200,000 weeks hrs, the number of defects per $1M in Scope and number of defects per 200,000 Work-hr parameters of the construction safety. Different optimization techniques are utilized to discover an optimal weight of the ANN process. All the optimum results demonstrate that the attained error values between the output of the experimental values and the predicted values are closely equal to zero in the designed network. From the results, the minimum error of 89.97% determined by the ANN is attained by the Grey Wolf Optimization (GWO) algorithm.