Cracking Analysis and Process Improvement of Stellite Alloy Overlay Welding on Steam-Driven Pump Balance Disc

Literature Overview

The paper by Liu Xing (2020), published in Nuclear Science and Engineering (Vol. 40, No. 4, pp. 718-728), addresses a critical reliability issue discovered during the overhaul of China's first domestically manufactured steam-driven feedwater pump. The balance disc, a key component responsible for axial force balancing in the pump rotor, exhibited cracking in its Stellite alloy overlay layer. This finding is particularly significant because steam-driven feedwater pumps are among the most safety-critical rotating equipment in nuclear power plants, and any degradation in the balance disc overlay layer directly threatens the integrity of the rotor assembly and, by extension, the nuclear safety system.

Core Findings and Root Cause Analysis

The investigation employed a systematic failure analysis methodology combining macroscopic observation, microstructural examination, energy-dispersive spectroscopy (EDS), and cross-sectional crack profiling. The key findings can be summarized as follows:

Crack morphology: The cracks originated at the fusion boundary between the base material and the Stellite overlay layer and propagated in both the overlay layer and the heat-affected zone (HAZ). The crack paths exhibited a mixture of intergranular and transgranular features, indicating a complex fracture mechanism.
EDS results: The composition analysis revealed localized enrichment of carbon and chromium at the crack initiation sites, suggesting the formation of brittle carbides (primarily Cr7C3 and Cr23C6) at the fusion interface.
Axial force calculation review: The rotor axial force calculation process was traced and found to contain assumptions that did not fully account for transient operating conditions, leading to underestimation of the axial load on the balance disc.
Overlay welding process deficiencies: The original welding procedure specification (WPS) did not adequately control interpass temperature, welding sequence, and post-weld heat treatment (PWHT) parameters, resulting in excessive residual stress accumulation.
Machining process deficiencies: The post-weld machining allowance was insufficient, and the machining sequence did not follow a progressive finishing strategy, leaving surface residual stresses unrelieved.

Technical Points and Process Improvement

The improvement measures implemented can be organized using a PDCA (Plan-Do-Check-Act) framework:

Process Stage	Original Condition	Improved Condition	Technical Rationale
Welding sequence	Single-direction, arbitrary start point	Symmetrical, spiral sequence from center outward	Reduces directional residual stress concentration
Interpass temperature	Not controlled	Strictly maintained at 150-250°C	Prevents excessive cooling rate and brittle phase formation
Preheating	Minimal or absent	250-350°C depending on wall thickness	Reduces thermal gradient and hydrogen-induced cracking susceptibility
Post-weld heat treatment	Inadequate or omitted	650-700°C for 2-4 hours followed by controlled cooling	Recrystallizes the HAZ, relieves residual stress, and dissolves brittle carbides
Machining allowance	0.5-1.0 mm	1.5-2.0 mm with progressive finishing	Removes surface defects and residual stresses from welding
Axial force design margin	Nominal operating condition	Includes transient and off-design conditions	Ensures adequate design margin under all credible operating scenarios

The improved overlay welding procedure also incorporated a multi-pass strategy with a transition layer between the base material and the Stellite alloy to mitigate the metallurgical incompatibility. The transition layer, typically composed of a low-carbon stainless steel or a nickel-based alloy, serves to reduce the carbon activity gradient and minimize carbide precipitation at the fusion boundary.

Engineering Practice Implications

This case study provides several important lessons for nuclear-grade rotating equipment maintenance and manufacturing:

Material compatibility is paramount: Stellite alloys (typically Stellite 6, with approximately 60% Cr, 28% Co, and 10% Ni) are known for their excellent wear and corrosion resistance but are inherently susceptible to cracking when applied to carbon steel or low-alloy steel substrates due to the large difference in thermal expansion coefficients and carbon activity. The thermal expansion coefficient mismatch can reach 2-3 × 10⁻⁶/°C, generating significant thermal residual stresses upon cooling.
Nuclear component qualification requires conservative design margins: The axial force calculation methodology must account for all credible operating conditions, including startup, shutdown, load rejection, and transient scenarios. The balance disc is a passive safety component, and its failure could lead to rotor seizure and subsequent loss of feedwater supply to the reactor coolant system.
Process validation through factory acceptance testing (FAT): The improved balance disc underwent comprehensive FAT, including dimensional inspection, non-destructive testing (NDT) with magnetic particle testing (MT) and penetrant testing (PT), hardness verification, and functional performance testing. The results demonstrated significantly improved resistance to crack initiation and propagation.
Standardization for future applications: The improved process has been incorporated into the design specifications for subsequent Hualong One (HPR1000) reactor units, providing a replicable engineering reference for domestic steam-driven pump manufacturing.

Study Insights and Reflections

This case exemplifies the importance of holistic failure analysis that integrates metallurgical investigation with mechanical design review and process engineering assessment. The root cause was not a single factor but a combination of design margin insufficiency, welding procedure inadequacy, and machining process deficiency. In my experience with nuclear-grade components, such multifactorial failures are common and demand a systems-thinking approach to resolution. The success of the improved process underscores the principle that overlay welding quality is determined not only by the welding operation itself but by the entire process chain from design through post-weld finishing. For engineers working on similar applications, I would emphasize the criticality of conducting a thorough weldability assessment before specifying overlay materials, and of implementing rigorous process controls at every stage of manufacturing.