Tenaris
Case Study: Reduction of quality failures in welding.
Challenge
Improve welding quality in the final stage of piping production for the Petrochemical industry.
Predict the occurrence of defects and point out the most influential variables in failures.
Optimize the parameters and propose adjustments to the process to maximize the quality of the final product and reduce costs.
N-Hence Manufacturing:
Statistical modeling to evaluate conformance vs. specifications and identify the probability of occurrence of defects in 7 welding stages.
Algoritmo de ML para calcular os parâmetros ótimos e ajustes no processo.
Online dashboard for parameter simulation and optimization.
Results:
All actions from the N-Hence Manufacturing solution generated excellent results for the company. Some example include:
- The model was able to anticipate 50% of defects based on 3 key variables.
- Decision based on data from the identification of probabilistic scenarios
- Immediate application of the results, (operator / parameters) in the welding process.
Impact
10% REDUCTION OF DEFECTS IN THE FIRST MONTH OF IMPLEMENTATION.