Water-droplet erosion behavior of high-velocity oxygen-fuel-sprayed coatings for steam turbine blades

The water-droplet erosion of low-pressure steam turbine blades under wet steam environments can alter the vibration characteristics of the blade, and lead to its premature failure. Using high-velocity oxygen-fuel (HVOF) sprayed water-droplet erosion resistant coating is beneficial in preventing the erosion failure, while the erosion behavior of such coatings is still not revealed so far. Here, we examined the water-droplet erosion resistance of Cr3C2–25NiCr and WC–10Co–4Cr HVOF sprayed coatings using a pulsed water jet device with different impingement angles. Combined with microscopic characterization, indentation, and adhesion tests, we found that: (1) both of the coatings exhibited a similar three-stage erosion behavior, from the formation of discrete erosion surface cavities and continuous grooves to the broadening and deepening of the groove, (2) the erosion rate accelerates with the increasing impingement angle of the water jet; besides, the impingement angle had a nonlinear effect on the cumulative mass loss, and 30° sample exhibited the smallest mass loss per unit area (3) an improvement in the interfacial adhesion strength, fracture toughness, and hardness of the coating enhanced the water-droplet erosion resistance. These results provide guidance pertaining to the engineering application of water erosion protective coatings on steam turbine blades.

Water Droplet Erosion Life Prediction Method for Steam Turbine Blade Materials Based on Image Recognition and Machine Learning

In this paper, a method for predicting remaining useful life (RUL) of turbine blade under water droplet erosion (WDE) based on image recognition and machine learning is presented. Using the experimental rig for testing the WDE characteristics of materials, the morphology pictures of specimen surface at different times in the process of WDE are collected. According to the data processing method of ASTM-G73 and the cumulative erosion-time curves, the WDE stages of materials is quantitatively divided and the WDE life coefficient (?) is defined. The life coefficient (?) could be used to calculate the RUL of turbine blades. One convolutional neural network model and three machine learning models are adopted to train and predict the image dataset. Then the training process and feature maps of the Resnet model are studied in detail. It is found that the highest prediction accuracy of the method proposed in this paper can be 0.949, which is considered acceptable to provide reference for turbine overhaul period and blade replacement time.

Experimental investigation on the water droplet erosion characteristics of blade materials for steam turbine

In this paper, the water droplet erosion ( WDE) characteristics of four blade materials under different high-speed liquid-solid impingement conditions are tested with the experimental steps and data processing methods in ASTM-G73. On the basis of cumulative erosion curves, the maximum erosion rate ( ERmax) is obtained to analyze the WDE resistance of the testing materials quantitatively. The effects of target surface roughness, impact angle and impact velocity on the WDE characteristics of blade materials under specific working conditions are studied. The velocity coefficients ( n) of the testing materials under high-speed liquid-solid impact are analyzed. In order to investigate the mechanism of WDE, the material failure characteristics and shallow layer hardness ( SLH) at three characteristic moments corresponding to different WDE periods, are investigated with the SEM morphology of erosion section and hardness distribution along the depth direction at the impact position. The research results can provide reference data and technical support for the structure design and material selection of steam turbine blades.

Experimental Research on Water Droplet Erosion Resistance Characteristics of Turbine Blade Substrate and Strengthened Layers Materials

In this paper, the water droplet erosion (WDE) performance of typical martensitic precipitation substrate 0Cr17Ni4Cu4Nb in steam turbine final stage, laser solid solution strengthened sample, laser cladding sample and brazed stellite alloy samples have been studied based on a high-speed rotating waterjet test system. The WDE resistance of several materials from strong to weak is in sequence: Brazed stellite alloy > laser cladding sample > laser solid solution sample > martensitic substrate. Furthermore, the WDE resistance mechanism and the failure mode of brazed stellite alloy have been revealed. It is found that the hard carbide in the stellite alloy is the starting point of crack formation and propagation. Under the continuous droplet impact, cracks grow and connect into networks, resulting in the removal of carbide precipitates and WDE damage. It is proved that the properties of the Co-based material itself is the reason for its excellent WDE resistance. And the carbides have almost no positive contribution to its anti-erodibility. These new findings are of great significance to process methods and parameter selection of steam turbine blade materials and surface strengthened layers.

Water Droplet Erosion Life Prediction Method for Steam Turbine Blade Materials Based on Image Recognition and Machine Learning

In this paper, a method for predicting remaining useful life (RUL) of turbine blade under water droplet erosion (WDE) based on image recognition and machine learning is presented. Using the experimental rig for testing the WDE characteristics of materials, the morphology pictures of specimen surface at different times in the process of WDE are collected. According to the data processing method of ASTM-G73 and the cumulative erosion-time curves, the WDE stages of materials is quantitatively divided and the WDE life coefficient (ζ) is defined. The life coefficient (ζ) could be used to calculate the RUL of turbine blades. One convolutional neural network model and three machine learning models are adopted to train and predict the image dataset. Then the training process and feature maps of the Resnet model are studied in detail. It is found that the highest prediction accuracy of the method proposed in this paper can be 0.949, which is considered acceptable to provide reference for turbine overhaul period and blade replacement time.