Denis Thibault

Impact of startup scheme on Francis runner life expectancy

Francis runners are subject to complex dynamic forces which might lead to eventual blade cracking and the need for corrective measure. Damage due to cracks in runner blades are usually not a safety issues but might generate unexpected down time and high repair cost. Avoiding the main damaging operating conditions is often the only option left to plant operators to maximize the life expectancy of their runner. The startup scheme is one of the available parameter which is controlled by the end user and could be used to minimize the damage induced to the runner. In this study, two startup schemes have been used to investigate life expectancy of Francis runner using in situ measurements. The results obtained show that the induced damage during the startup event could be significantly reduced with change to the startup scheme. In our opinion, an optimization of the startup scheme with regard to fatigue damage could extend significantly the life expectancy and the reliability of Francis runner.

An Open-Source Engine for the Processing of Electron Backscatter Patterns: EBSD-Image

An open source software package dedicated to processing stored electron backscatter patterns is presented. The package gives users full control over the type and order of operations that are performed on electron backscatter diffraction (EBSD) patterns as well as the results obtained. The current version of EBSD-Image (www.ebsd-image.org) offers a flexible and structured interface to calculate various quality metrics over large datasets. It includes unique features such as practical file formats for storing diffraction patterns and analysis results, stitching of mappings with automatic reorganization of their diffraction patterns, and routines for processing data on a distributed computer grid. Implementations of the algorithms used in the software are described and benchmarked using simulated diffraction patterns. Using those simulated EBSD patterns, the detection of Kikuchi bands in EBSD-Image was found to be comparable to commercially available EBSD systems. In addition, 24 quality metrics were evaluated based on the ability to assess the level of deformation in two samples (copper and iron) deformed using 220 grit SiC grinding paper. Fourteen metrics were able to properly measure the deformation gradient of the samples.

New parameters influencing hydraulic runner lifetime

Traditionally, hydraulic runner mechanical design is based on calculation of static stresses. Today, validation of hydraulic runner design in terms of reliability requires taking into account the fatigue effect of dynamics loads. A damage tolerant approach based on fracture mechanics is the method chosen by Alstom and Hydro-Quebec to study fatigue damage in runners. This requires a careful examination of all factors influencing material fatigue behavior. Such material behavior depends mainly on the chemical composition, microstructure and thermal history of the component, and on the resulting residual stresses. Measurement of fracture mechanics properties of various steels have demonstrated that runner lifetime can be significantly altered by differences in the manufacturing process, although remaining in accordance with agreed practices and standards such as ASTM. Carbon content and heat treatment are suspected to influence fatigue lifetime. This will have to be investigated by continuing the current research.

The role of high cycle fatigue (HCF) onset in Francis runner reliability

High Cycle Fatigue (HCF) plays an important role in Francis runner reliability. This paper presents a model in which reliability is defined as the probability of not exceeding a threshold above which HCF contributes to crack propagation. In the context of combined Low Cycle Fatigue (LCF) and HCF loading, the Kitagawa diagram is used as the limit state threshold for reliability. The reliability problem is solved using First-Order Reliability Methods (FORM). A study case is proposed using in situ measured strains and operational data. All the parameters of the reliability problem are based either on observed data or on typical design specifications. From the results obtained, we observed that the uncertainty around the defect size and the HCF stress range play an important role in reliability. At the same time, we observed that expected values for the LCF stress range and the number of LCF cycles have a significant influence on life assessment, but the uncertainty around these values could be neglected in the reliability assessment.

Residual Stresses Induced by Robotized Hammer-Peening

The strains induced by hammer-peening were measured by strain gauges on a mild steel plate during the hammer-peening operation. This process has recently been robotized by the research institute of Hydro-Québec (IREQ), so the hammer-peening performed for this study was done with the help of a Scompi robot. The resulting stresses calculated from the strain measurements were compared with residual stress measurements made with the hole drilling technique. The comparison shows a very good correlation of the two sets of measurements. Residual stress measurements were also made in the hammer-peened zone: as expected we found a highly biaxial compressive state of stress in this zone. An unexpected region of transverse tension was found at the end of the hammerpeened zone. This region can be very critical if hammer-peening is made with the objective of improving fatigue behaviour.

An Experimental Comparison of Weld-Induced Residual Stresses Using Different Stainless Steel Filler Metals Commonly Used for Hydraulic Turbines Manufacturing and Repair

CA6NM and UNS S41500 martensitic stainless steels are widely used for manufacturing and repair of hydraulic turbine runners. They offer good mechanical properties and superior cavitation resistance when compared to mild steels. They are also relatively easy to weld. However, when welded homogeneously, they require a post-weld heat treatment (PWHT) in order to temper the as-welded martensite. This PWHT is also beneficial for residual stresses reduction as it effectively lowers the stress peaks. To avoid this PWHT, austenitic filler metals are often used for repair. But omitting PWHT inevitably leaves weld-induced residual stresses in the assembly. In order to better understand the impact of the weld filler metal choice on the importance of residual stress, an experimental study has been conducted on three different filler alloys. The chosen alloys were: • 410NiMo, a martensitic grade having the same composition as the base metal (13%Cr-4%Ni-0.5%Mo) ; • 309L, an austenitic grade widely used for repair (24%Cr-13%Ni) ; • A proprietary low transformation temperature (LTT) martensitic grade (13%Cr-6%Ni). This paper compares residual stresses in the as-welded condition on welds of UNS S41500 (13%Cr-4%Ni) made using these filler metals. Residual stresses were measured using the contour method. Microstructural analysis was performed to identify the phases in the weld and the heat-affected zones (HAZ). Microhardness maps were done to see the hardness distribution of each weldment.

Hydraulic Runner Design Method for Lifetime

Quest for reliability of hydraulic runners is a concern for all mature electricity producers. The fatigue damage caused by dynamics loads is frequently the root cause of runner failure. This paper presents the damage tolerance approach based on fracture mechanics as the method chosen by Alstom and Hydro-Quebec to predict effects of damage on runner lifetime and consequently to be use as a design method. This is sustained by a research on fracture mechanics properties of runner materials and by recommendations on the strategy to define a safety margin for design. The acquired knowledge permits to identify potential improvement of the runner lifetime without significant cost increase, like being more specific on some chemical composition or heat treatment.

Assessment of cold cracking tests for low transformation temperature martensitic stainless steel multipass welds

In the last decades, low-carbon soft‐martensitic stainless steels have been widely used by heavy industries such as hydraulic turbine and pipeline manufacturing. Their good corrosion resistance, high mechanical strength, and good cavitation erosion resistance combined with their relatively good weldability explain this success. However, arc welding operations of thick plates still require significant preheating to avoid cold cracking. From a practical point of view, this requirement can complicate manufacturing operations and limit the use of these alloys for some applications. It thus becomes important to assess their cold cracking susceptibility in order to reduce preheating temperature. The discrepancy between industrial observations in multipass welds and the results obtained from the standardized cold cracking test such as Tekken and GBOP (gapped bead-on-plate) tests led to a modification of the GBOP test as presented here. In order to reproduce the behavior of industrial cold cracks, a comparison between Tekken and GBOP tests was set up for the assessment of the cold cracking susceptibility of multipass 13%Cr‐4%Ni soft martensitic stainless steels welded with 410NiMo filler alloy. A two‐bead GBOP test was proposed. This modification is necessary because the low martensitic transformation temperature of the 13%Cr‐4%Ni stainless steels reverses the residual stress in the weld, preventing the possibility of cold cracking. A second bead is needed to induce tensile stress in the first bead so that crack may be generated as in industrial conditions. The relevance of these modifications is presented and discussed using contour method residual stress measurements and acoustic emission crack detection.

The effect of materials properties on the reliability of hydraulic turbine runners

The failure of hydraulic turbine runners is a rare event. So in order to assess the reliability of these components one cannot rely solely on the number of observed failures in a given population. However, as there is a limited number of degradation mechanisms involved, it is possible to use physically-based reliability models. Such models are often more complicated but are able to account for physical parameters in the degradation process. They can therefore help provide solutions to improve reliability. With such models, the effect of materials properties on runner reliability can be highlighted. This paper presents a brief review of the Kitagawa-Takahashi diagram which links the damage tolerance approach, based on fracture mechanics, to the stress or strainlife approaches. Using simplified response spectra based on runner stress measurements, we will show how fatigue reliability is sensitive to materials fatigue properties, namely fatigue crack propagation behaviour and fatigue limit obtained on S-N curves. Furthermore, we will review the influence of the main microstructural features observed in 13%Cr-4%Ni stainless steels commonly used for runner manufacturing. The goal is ultimately to identify the most influential microstructural features and to quantify their effect on fatigue reliability of runners.

Methodology for estimating strain gauge measurement biases and uncertainties on isotropic materials

Compared to actual strains, the values obtained with strain gauges during experimental measurements contain biases and uncertainties. In this article, we propose a methodology using Monte Carlo simulations to estimate the effects of biases and uncertainties from the following: location uncertainty, integration effect and transverse sensitivity errors in unidirectional strain gauges. Moreover, the specific behaviour of welded gauges is also considered. The approach simulates strain gauges on the displacement fields obtained from the structure’s finite element analyses to predict the expected biases and uncertainties. With the use of experimental measurements designed to highlight the biases between gauge measurements and finite element analyses strain results, we verify the methodology. In our experimental verification, we observe that biases are adequately predicted by the proposed method. It is worth mentioning that such an approach can be used not only for validations between finite element analyses and experimental measurements but also for optimizations of strain gauge positioning during measurement campaigns.