Investigation of Learning Curve Model for Piling Work

          Learning Curves are taken in as an important factor that generally points out the variation in the productivity, which usually is counted in the construction projects during the estimation and planning step. In this study, it is set out to look upon the feasibility of Learning Curve models for the analysis of learning experience using productivity data in the piling process. The same set of data will always bring out different learning rates under the applied models because unit time and cumulative average time do not dip down at the same rate. The best model is the one that predicts the numbers or other outputs that are closest to the actual results in terms of time or cost estimation. In this study, the role of different LC models (i.e., Wright, Quadratic and Hyperbolic) is looked into by the comparison of the outcomes by way of the application of cumulative productivity data of the activity. The main objectives are the investigation of the model which is the best model for historical data and finding out the best model that brings out the better forecast. The best model is decided based on the least deviation from the yielded results of each model with respect to the actual data. Analysis of the cumulative average productivity data predicted that the Wright Model beats all models both the cases of fitting Historical data and in predicting future performance.

          In this Independent Study, the author was planning to gather productivity data of rectangularpiles, size 0.22 x 0.22 x 6.00 m. The data will be grouped and analyzed by way of mathematical modelling for the learning curve of the piling work which is a substructure component of the flood wall structure of the raw water supply system improvement project to enhance stability in the west canal area from the Mahasawat Water Plant to the Bang Lane Raw Water Pumping Plant. It is expected that the outcomes of the study will be useful in terms of cost analysis for job bidding and planning for a number of projects in the future.