2500 KW Ship Service Turbine Generator Casing Welded Inconel Plug Failure and Repair Analysis

Abstract

Six of seven US Navy Wasp Class (LHD) ships have 2500 KW Ship Service Turbo Generator (SSTG) steam turbine casings that were mistakenly manufactured with in-place first stage balancing holes in both the inboard and outboard sides of the upper casing half. These holes were intended for in-place balancing of the turbine rotor; however, the US Navy did not request the access holes since in-place balancing is typically not accomplished on in-service surface ship SSTGs. To correct the mistake, the OEM, developed a procedure to weld plugs into the holes. Unfortunately, cracks developed in the plug weld heat affected zones (HAZs) on many of the in-service units. Some cracks propagated entirely through the plug and leaked steam in service. A failure analysis determined that the original plug (Inconel X750) should have received post weld heat treatment (PWHT) to avoid embrittlement and the subsequent cracking of the HAZ. If PWHT was not to be accomplished, an alloy such as Inconel 600 should have been selected. It is noted that PWHT risks warping the casing and cannot be performed in-place. Inconel 600 repair plug installations were performed in-place, permitting a fast repair turnaround time, which allowed the affected ships to meet operational schedules. To evaluate the repair integrity, the US Navy reviewed the failure analysis data and repair procedure and performed a Finite Element Analysis (FEA) and Fracture Analysis. Various “improved plug” designs were also studied in parallel to determine if a different geometry plug would better resist future cracking. During repairs, minor radial cracks were discovered in the welding “inlay” originally installed between the casing and plug weld. Because the inlay had to be fully preserved to avoid complex and time consuming additional repairs to the casing including PWHT, minor cracks were left in place and consumed by the new weld. An important objective of this effort was to prove that if any cracking ever reoccurred only minor steam leakage would result. The leakage would be apparent to the operators long before there was potential for the plug to fail catastrophically; moreover, a “leak-before-break” determination was requested to be proven and validated.