Prediction of tensile failure load for maraging steel weldment by acoustic emission technique

Abstract

Maraging steel (Grade 250) pressurized chambers are used in booster stages for the launch vehicles and missiles. These are designed & realized with ultra high strength steels like Maraging steels in order to gain in range and pay load capabilities with optimal Factor-of-Safety(FOS). Effective manufacturing methodology and fracture control is a prime requirement however manufacturing processes give rise to defects that are inevitable. Non-Destructive Engineering techniques viz. ultrasonic testing, Radiography testing, Eddy current testing etc. address the issue of detecting passive defects only. Defects that are active i.e. critical and that could cause catastrophe shall be detected possibly during Proof pressure testing Strain gauge method adopted is restricted for strain measurements and is in-effective in detecting flaws as it is a highly localized. Acoustic Emission Technique (AET), as a whole field method, has ability to detect unstable flaws effectively. Acoustic emissions are generated due to defects, microstructural variation, presence of inclusions and second phase particles in metallic materials. AET identifies defects and discontinuities in terms of Acoustic Emission parameters. Sources of Acoustic Emission (AE) can be distinguished by their AE signature in terms of amplitude, Count and Energy. The severity can be quantified in terms of AE Parameters. In this paper an attempt is made towards predicting tensile failure load of Maraging Steel weldment with varying extents of notch thereby representing equivalent tight cracks as per Linear Elastic Fracture Mechanics (LEFM) design approach. Customized specimens were fabricated and notches were made using Electric Discharge Machining process. Tensile load has been applied to the test specimens with AE data acquisition. The AE distribution obtained from each specimen has been correlated to an equivalent Weibull distribution and represented in terms of weibull parameters. The significant Weibull parameters viz. Skewness (b value) and centroid of distribution curve (θ) are estimated. The distribution at a load of 85% of failure load is used for the prediction process. An empirical relation connecting Weibull parameters b, θ, b * θ, is proposed. It is observed that the product b * θ is linearly correlated to tensile failure load. On comparing results, predicted tensile failure load is closely matching to the recorded tensile failure load of the specimen. The average prediction capability of the proposed model is within 5–6%.