Predicting the effects of surface treatments

By Dr. N. Jayaraman

The term “lifing” generally refers to making predictions about the service life of components. Lifing analysis uses existing material properties and models to predict the fatigue or service life of critical components when subjected to various conditions.

This process is crucial at early stages of component design to prevent malfunctions that could lead to failure. It’s also frequently used in revision of existing components, allowing adoption of a new solution when failures occur.

Estimating the service life of a component is the main benefit of lifing analysis. Considering the high cost of most turbomachinery components, testing them with multiple surface treatments or different conditions to achieve optimal service life is usually cost-prohibitive. This is because reliable testing to prove-out these different methods is generally destructive, like full-scale component level fatigue testing or residual stress measurement.

If a component is designed for a targeted service life, other useful predictions from lifing analysis include:

• Determining the appropriate operating conditions to yield the optimum service life
• Calculating the expected damage tolerance improvement
• Establishing the expected service life of the component under various operating conditions

How it Works

Depending on the design requirements, lifing analysis can be relatively simple or quite involved. This is dependent on the component operating conditions. Relatively fixed operating conditions only require relatively simple analysis, while more complex conditions warrant a more detailed analysis. This includes material behavior data.

Common materials with well-known properties are easy candidates for lifing. In cases where rarer materials are used, a proprietary or new alloy for example, the availability of other information is crucial. These include:

• Tensile: Elastic modulus, Poisson’s ratio, yield strength, ultimate tensile strength, strain hardening parameters
• Fatigue: Stress /Strain – life data, available at different R-ratios (R=min stress/max stress)
• Fatigue crack growth data

The design of the component is also important, particularly stress/strain distributions in critical regions. All of these factors are required to put together a reliable prediction of component life.

These factors are applied to models like the fatigue design diagram (FDD) to estimate component life or fracture mechanics-based methods to estimate crack growth. Results from both of these approaches combined are complimentary, providing a clear path to a successful lifing solution for the component.

Lifing Analysis in Action

One lifing analysis program undertaken at Lambda Technologies Group explored life extension options for the 7FA R0 gas turbine blades. The proposed surface treatment was low plasticity burnishing (LPB), a highly controlled mechanical process that applies engineered residual compression to the critical areas of components.

As reported by EPRI, the life of the 7FA gas turbine is about 144,000 hours, or 5,000 starts. The primary life-limiting factor of the R0 compressor blades is high cycle fatigue cracking in the leading edge. The primary cause of crack initiation has been attributed to erosion and inherent low damage tolerance for the component.

The goal of this project was to reduce costs. The frequent erosion inspections, blade replacement interval requirements, and necessary downtime required for each would be decreased dramatically by extending the component life of the blades.

Using fracture mechanics based lifing analysis and our proprietary Fatigue Design Diagram (FDD), a modification of the Haigh diagram method, it was possible to determine the minimum residual compression required to mitigate high cycle fatigue cracking from erosion damage in the R0 blade. The models predicted exponential life extension.

The pre-determined field of residual compression was then applied to the blades using the LPB process. Component level high cycle fatigue testing was performed to confirm the success of the application. EDM notches were used to simulate deep erosion damage.

Fatigue testing showed that the applied residual compression mitigated fatigue cracking from deep erosion damage in this application. Baseline (unprocessed) blades with 0.008 inch deep damage – the allowed maximum before repair or replacement – all failed in under 150,000 fatigue cycles. The LPB processed samples did not fail, running out to more than 15 million cycles, even with 0.025 inch deep damage.

The cost of this lifing analysis was in the low four-figures range and took less than two weeks to complete. It enabled the company to put funding into their chosen surface treatment without worry that it might not be the right solution for its application.

Results of the program showed a damage tolerance improvement of over 3x in the R0 blade. Conservatively, the 3x damage tolerance increase could reduce the inspection and blending requirement of the R0 blades to about 1/3 of the current frequency, resulting in a corresponding decrease in cost to approximately 1/3 of those associated with downtime, inspection, repair and replacement.

Dr. N. Jayaraman is the Director of Materials Research for the Lambda Technologies Group, a company focusing on materials testing and life extension technologies. For more information, call +1 (800) 883-0851 or visit lambdatechs.com.