Search results

The purpose of the study is to analyse the feasibility of using the potential and exponential curve models to assess the learning of a group of welders, when welding stainless steel piping with the tungsten inert gas process.

The welding productivity data grouped according to the requirements of the ASME SECTION IX code is organised into two groups: average productivity and baseline productivity. When processing the adjustment to the two models, the Excel software Solver tool was used. The criteria for assessing the quality of the fit were: least squared method, Spearman's correlation coefficient and graphical method. The impact of the variation coefficient on the average productivity and the amplitude (difference between the minimum and maximum productivity) was also evaluated on the baseline productivity.

The curves elaborated based on the average productivity presented better quality of adjustment than those constructed from the baseline productivity. The potential and exponential models presented similar adjustment conditions, with the second having a slightly superior performance. There were no productivity gains due to learning in the studied time interval. The grouping of the average daily productivity data based on the diameter range established in the ASME code section IX presented satisfactory results, enabling its use by the industry.

There is no news of work on piping welding with this focus. The proposal to group the productivity data according to the degree of difficulty of execution established by the ASME code section IX, widely used in the industry, is a significant contribution to monitoring the evolution of learning. In the same way, the results allow to adopt the average productivity determined from the first 20 days of realisation of a project, as a reasonable indicator to estimate the future performance of the work, helping to correct deadlines during the realisation of a project.