Laser additive cladding method and system for repairing defects of steam turbine blade

Abstract

The application provides a turbine blade laser additive cladding defect repairing method and a regulation system, relates to the technical field of laser additive remanufacturing, and comprises the following steps: grain structure defect mechanism analysis and ordered growth regulation; type thermal crack mechanism identification and inhibition regulation; pore defect classification identification and accurate control; integrated defect closed-loop repair. Based on the mechanism model and regulation criterion of three typical defects of grain structure, thermal crack and pore, the whole set of laser additive cladding process parameters are optimized and matched, and the defect-free accurate repair of the damaged area of the turbine blade is completed. The application can realize the full-dimension accurate control of typical defects of grain structure, crack and pore in the laser additive cladding process of the turbine blade, greatly improves the repair forming quality, mechanical properties and service life of the blade, and is suitable for efficient non-destructive repair of damaged blades of various thermal power and nuclear power turbine, and has strong engineering applicability.

Description

Technical Field

[0001] This invention relates to the fields of laser additive remanufacturing and turbine equipment repair technology, and particularly to a method and system for repairing defects in turbine blades using laser additive cladding. Background Technology

[0002] Steam turbines are one of the core power units of modern energy equipment and a key foundation for ensuring national energy security and efficient supply. Steam turbine blades operate under the combined effects of high-temperature, high-humidity steam, complex aerodynamic loads, and strong centrifugal forces for extended periods, making them susceptible to wear, corrosion (including water droplet / solid particle erosion and stress corrosion cracking), thermal fatigue, and high-cycle fatigue cracking, leading to decreased efficiency and reliability degradation. Large equipment manufacturers in Europe and America have taken the lead in developing the blade remanufacturing and additive repair industry chain, forming a systematic process and service capability for repairing key components of large military and civilian steam turbines, and mastering high-energy beam additive repair production lines represented by laser-directed energy deposition (DED) / laser cladding. Due to the influence of key processes and quality assurance systems, related technologies have long been monopolized and face technological barriers in the high-end power equipment field, becoming one of the key links restricting my country's industrial upgrading and the improvement of high-end operation and maintenance service capabilities. Steam turbine blades are mostly made of high-strength, tough, and corrosion-resistant martensitic stainless steel, precipitation-hardening stainless steel, or nickel-based alloys, with complex manufacturing and heat treatment processes, high costs, large scale, and huge investment. Scrapping blades with only localized wear, edge cracks, or minor casting / machining defects and replacing them with new ones results in a significant waste of materials and manufacturing capacity. Furthermore, downtime for maintenance and spare parts supply is costly and time-consuming. Developing efficient, high-precision, and high-quality laser additive repair and remanufacturing technologies to restore blade geometry and surface integrity, and to reconstruct microstructure and mechanical / corrosion-resistant properties that match the base material, thus meeting re-service requirements, is crucial for improving unit economy, reducing life-cycle maintenance costs, and enhancing equipment reliability and safety redundancy. However, due to the complex surface structure, special microstructure, and harsh operating load conditions of turbine blades, implementing high-energy beam repair not only requires consideration of restoring blade geometry and reducing metallurgical defects in the repair area, but also presents a greater challenge in the manufacturing and in-situ repair process, which has the following characteristics: ① The process is physically demanding and the process window is complexly coupled: the synergy of multiple fields—laser-powder flow-matrix—leads to a high-gradient temperature field and rapid solidification behavior, which can easily induce defects such as keyholes, incomplete welding, microcracks, inclusions, and spheroidization; ② The sensing data is multi-source, heterogeneous, and highly interfering: parallel acquisition of multiple modes, including high-speed visible / infrared imaging, emission spectroscopy, acoustic emission / laser ultrasound, backlight / reflection power, coaxial vision, and OCT / laser displacement, results in non-stationary noise, cross-frequency coupling modulation, asynchronous sampling, and viewpoint obstruction, leading to significant spatiotemporal heterogeneity; ③ Complex free-form surfaces and local obstruction make energy coupling, thermal accumulation, and residual stress more sensitive, with mutual constraints between dense forming, low cracking, and microstructure performance matching. Summary of the Invention

[0003] The purpose of this invention is to provide a method and system for repairing defects in steam turbine blades using laser additive cladding.

[0004] A method for repairing defects in steam turbine blades using laser additive cladding includes the following steps: Step 1: Analysis of Grain Structure Defect Mechanism and Orderly Growth Control: For martensitic stainless steel and heat-resistant steel substrates for turbine blades, the temperature gradient G of the molten pool and the solid-liquid interface velocity R during the laser melting deposition process were collected. A critical mathematical model of columnar-equiaxed crystal transformation coupled with GR was established to analyze the grain orientation instability law. Combining heterogeneous nucleation kinetics, solute element redistribution mechanism and molten pool Marangoni convection effect, the formation causes of grain skewness and orientation mixing were clarified. By matching the GR parameter threshold, the ordered growth of grain structure in the deposition layer was achieved. Step 2: Identification and Suppression of Multiple Types of Hot Cracks: For the high-constraint, multi-pass, multi-layer laser deposition repair scenario of blades, a thermo-elastic-plastic-phase-transition multi-field coupled simulation model is constructed. Combined with the micro-morphological characterization of cracks, three types of thermal defects are distinguished: solidification cracks, liquefaction cracks, and reheat cracks. The effects of dilution rate, multi-layer thermal cycling superposition, and distribution of material weak areas on stress concentration in the repair area are clarified. The continuous liquid film backfilling molding conditions are defined, a quantitative criterion for the tendency of hot cracking in laser additive repair is established, and corresponding process suppression paths are formulated. Step 3, Porosity Defect Classification, Identification and Precise Control: An integrated detection method combining in-situ high-speed imaging, infrared monitoring and optical microscopy is adopted to accurately distinguish between keyhole-type pores and non-fused pore defects; the influence of laser power, scanning speed, remelting strategy, scanning overlap rate, powder / wire feeding rate, protective gas purity and environmental moisture content on pore initiation, growth, closure and extinction is quantified, and an engineering pore formation mechanism model and standardized control criteria are constructed; Step 4, Integrated Defect Closed-Loop Repair: Based on the mechanism models and control criteria of three typical defects—grain structure, hot cracks, and porosity—the complete set of laser additive cladding process parameters is optimized and matched to complete the defect-free and precise repair of the damaged area of ​​the turbine blade.

[0005] Preferably, the critical mathematical model for the transformation from columnar to equiaxed crystals described in step one uses the grain growth undercooling as the core variable to limit the critical G / R ratio range for the transformation from columnar to equiaxed crystals. When G / R is higher than the upper critical value, the molten pool is dominated by directional columnar crystal growth; when G / R is lower than the lower critical value, a large number of heterogeneous nuclei are triggered, achieving uniform precipitation of equiaxed crystals.

[0006] Preferably, the grain orientation instability law described in step one is as follows: During the rapid solidification process of laser deposition, the Marangoni convection in the molten pool causes disordered flow of the melt, which destroys the stability of the solid-liquid interface. At the same time, the segregation of solute elements causes local nucleation barrier differences, which ultimately leads to the skewed grain growth direction and disordered mixed grain orientation. By controlling the convection intensity of the molten pool and stabilizing the interface temperature gradient, grain orientation defects can be eliminated.

[0007] Preferably, the three main controlling factors of hot cracks in step two are as follows: solidification cracks are caused by the enrichment of low-melting-point solute eutectic phases and the tensile stress caused by the contraction of continuous liquid film between grains at the end of deposition; liquefaction cracks are caused by the remelting of low-melting-point phases in the heat-affected zone of the substrate due to multi-layer thermal cycling, and the cracking of the liquid film between grains under the action of restraint stress; reheat cracks are caused by the precipitation of brittle phases at grain boundaries and the release of residual stress during high-temperature tempering after deposition, leading to grain boundary cracking.

[0008] Preferably, the quantitative criterion for thermal cracking tendency in step two is constructed based on the coupling of multiple parameters such as dilution rate threshold, number of thermal cycles, width of vulnerable area, and peak interface stress. When the calculated thermal cracking index is lower than the safety threshold, it is determined that there is no risk of thermal crack initiation in the repair area. The suppression path includes four core control methods: substrate preheating and temperature control, layer-by-layer cooling and pressure holding, gradient dilution rate process matching, and multi-layer scanning trajectory optimization.

[0009] Preferably, the classification and identification criteria for porosity defects in step three are as follows: Keyhole type pores are irregular large-diameter deep holes, caused by excessive laser energy leading to keyhole collapse of the molten pool and insufficient molten backfilling; Incomplete fusion type pores are small, diffuse micropores, caused by insufficient melting of powder and wire, insufficient interlayer overlap, and inclusion of protective gas atmosphere.

[0010] Preferably, the process parameter quantification and control rules in step three include: increasing laser power and reducing scanning speed can weaken unfused porosity defects, while excessively high power will exacerbate keyhole formation; adopting single-layer remelting and multi-layer gradient overlapping strategies can significantly close interlayer residual pores; strictly controlling the purity of protective gas to ≥99.99% and the ambient humidity to <30%RH can prevent the formation of gas-included pores.

[0011] Preferably, the range of optimized process parameters in step four is as follows: laser power 800-1800W, scanning speed 5-15mm / s, overlap rate 30%-50%, powder feeding rate 8-20g / min, substrate preheating temperature 200-400℃, and single-layer remelting times 1-2 times.

[0012] A control system for laser additive cladding repair of defects in steam turbine blades includes: a molten pool parameter monitoring module, a microstructure characterization module, a multi-field coupled simulation module, a defect intelligent identification module, a process parameter control module, and a laser additive execution module. The molten pool parameter monitoring module collects molten pool temperature, morphology, and environmental parameters in real time and transmits the data to the defect intelligent identification module for online real-time defect detection. The detected data is then transmitted to the multi-field coupled simulation module to construct dynamic temperature and stress field models. Metallographic defect data detected offline by the microstructure characterization module is synchronized to the multi-field coupled simulation module and the defect intelligent identification module to optimize the mathematical model of the defect mechanism. The multi-field coupled simulation module calculates the critical range of grain transformation, the hot cracking tendency index, and the porosity evolution law based on the real-time acquired data. The defect intelligent identification module simultaneously outputs the real-time defect type and location. Simulation results and defect early warning information are transmitted via OPC. The UA protocol is transmitted to the process parameter control module; the process parameter control module automatically optimizes the entire set of process parameters according to the defect control criteria, and sends control commands through the Profinet bus to drive the laser additive manufacturing execution module to adjust the laser, motion, powder feeding, and gas path systems; the laser additive manufacturing execution module transmits the actual operating status of the equipment back to the process parameter control module in real time, completing the real-time closed-loop communication of the entire system data acquisition, simulation identification, parameter control, and equipment execution.

[0013] Preferably, the molten pool parameter monitoring module is used to collect molten pool temperature gradient, interface growth rate, and molten pool convection state parameters in real time; the defect intelligent identification module can automatically distinguish grain defects, hot cracks, and two types of porosity defects, and match the corresponding process control scheme.

[0014] Using this invention, for the first time, a GR-coupled grain transformation critical model was established for martensitic stainless steel and heat-resistant steel materials of turbine blades. This model elucidates the influence mechanisms of heterogeneous nucleation, solute redistribution, and Marangoni convection on grain defects, fundamentally revealing the causes of grain orientation disorder and filling the mechanistic gap in the precise control of grain structure during laser repair of turbine blades. Furthermore, it distinguishes the differentiated main control mechanisms of solidification, liquefaction, and reheating hot cracks, constructing a thermo-elastic-plastic phase transformation coupled analysis model and a quantitative hot crack tendency criterion. This breaks away from the traditional empirical process control mode and enables high-constraint multilayer deposition scenarios. The technology enables precise prediction and suppression of hot cracks; it innovatively distinguishes between keyhole-type and non-fusion-type porosity defects, quantifies the influence of all core process parameters on porosity evolution, establishes engineering control criteria, and significantly reduces the residual rate of porosity defects in the repair layer; the entire technology forms a closed-loop technical system from defect mechanism, quantitative criteria, process criteria to forming repair, with quantifiable parameters and strong replicability, applicable to the repair of turbine blades of different damage degrees and specifications, and can significantly improve the blade repair qualification rate and service life, adapting to the needs of industrialized mass remanufacturing production. Attached Figure Description

[0015] Figure 1 A schematic diagram of the process for repairing defects in 30Cr13 martensitic stainless steel turbine blades using laser additive cladding. Figure 2 This is a schematic diagram of the macroscopic morphology of a single cladding layer. Figure 3 A schematic diagram of the structure for studying multi-layer lap cladding process; Figure 4 This is a schematic diagram of a typical additive manufacturing and joining process for thin-walled complex components. Figure 5 (a) Schematic diagram of numerical simulation of stress field in temperature field of additive manufacturing; (b) Schematic diagram of residual stress measurement by drilling method; Figure 6 A schematic diagram of the control system for repairing defects in steam turbine blades using laser additive cladding. The diagram shows: 1. Molten pool parameter monitoring module; 2. Microstructure characterization module; 3. Multi-field coupling simulation module; 4. Defect intelligent identification module; 5. Process parameter control module; 6. Laser additive manufacturing execution module. Detailed Implementation

[0016] To make the technical solution of the present invention easier to understand, the technical solution of the present invention will now be clearly and completely described in conjunction with the accompanying drawings and specific embodiments.

[0017] Example 1, please also refer to Figure 1 A method for repairing defects in steam turbine blades using laser additive cladding includes the following steps: Step S1, Analysis of Grain Structure Defect Mechanism and Orderly Growth Control: For martensitic stainless steel and heat-resistant steel substrates for turbine blades, the temperature gradient G of the molten pool and the solid-liquid interface velocity R during the laser melting deposition process are collected. A critical mathematical model of the columnar-equiaxed crystal transformation coupled with GR is established to analyze the grain orientation instability law. Combining heterogeneous nucleation kinetics, solute element redistribution mechanism and molten pool Marangoni convection effect, the causes of grain skewness and mixed orientation are clarified. By matching the GR parameter threshold, the ordered growth of grain structure in the deposition layer is achieved. The critical mathematical model of the columnar-equiaxed crystal transformation coupled with GR uses the grain growth undercooling as the core variable to limit the critical G / R ratio range of the transformation from columnar to equiaxed crystals. When G / R is higher than the upper limit of the critical value, the molten pool is dominated by directional columnar crystal growth. When G / R is lower than the lower limit of the critical value, a large number of heterogeneous nuclei are triggered, and equiaxed crystals are uniformly precipitated. The grain orientation instability is specifically described as follows: During the rapid solidification process of laser deposition, Marangoni convection in the molten pool causes disordered melt flow, which disrupts the stability of the solid-liquid interface. At the same time, solute element segregation causes local nucleation barrier differences, ultimately leading to skewed grain growth direction and disordered mixed grain orientation. By controlling the molten pool convection intensity and stabilizing the interface temperature gradient, grain orientation defects can be eliminated.

[0018] Step S2, Identification and Suppression of Multiple Types of Hot Cracks: For the high-constraint, multi-layer laser deposition repair scenario of blades, a thermo-elastic-plastic-phase transition multi-field coupled simulation model is constructed. Combined with the microscopic morphology characterization of cracks, three types of thermal defects are distinguished: solidification cracks, liquefaction cracks, and reheat cracks. The influence of dilution rate, multi-layer thermal cycling superposition, and distribution of material weak areas on stress concentration in the repair area is clarified. The continuous liquid film backfilling molding conditions are defined, a quantitative criterion for the tendency of hot cracking in laser additive repair is established, and corresponding process suppression paths are formulated. The main controlling factors of the three types of hot cracks are as follows: solidification cracks are caused by the enrichment of low-melting-point solute eutectic phase at the end of deposition and the tensile stress caused by the contraction of continuous liquid film between grains; liquefaction cracks are caused by the remelting of low-melting-point phases in the heat-affected zone of the substrate caused by multi-layer thermal cycling, and the cracking of the intergranular liquid film under the action of constraint stress; reheat cracks are caused by the precipitation of brittle phases at grain boundaries and the release of residual stress during the high-temperature tempering process after deposition, leading to grain boundary cracking. The quantitative criterion for thermal cracking tendency is constructed based on the coupling of multiple parameters such as dilution rate threshold, number of thermal cycles, width of vulnerable area, and peak interface stress. When the calculated thermal cracking index is lower than the safety threshold, it is determined that there is no risk of thermal crack initiation in the repair area. The suppression path includes four core control methods: substrate preheating and temperature control, layer-by-layer cooling and pressure holding, gradient dilution rate process matching, and multi-layer scanning trajectory optimization.

[0019] Step S3, Pore Defect Classification, Identification, and Precise Control: An integrated detection method combining in-situ high-speed imaging, infrared monitoring, and optical microscopy is employed to accurately distinguish between keyhole-type pores and non-fused pores. The influence of laser power, scanning speed, remelting strategy, scanning overlap rate, powder / wire feeding rate, protective gas purity, and environmental moisture content on pore initiation, growth, closure, and demise is quantified, constructing an engineering-based pore formation mechanism model and standardized control criteria. The pore defect classification and identification criteria are as follows: Keyhole-type pores are irregular, large-diameter, deep holes caused by excessive laser energy leading to keyhole collapse in the molten pool and insufficient melt backfilling; non-fused pores are small, dispersed micropores caused by insufficient powder / wire melting, inadequate interlayer overlap, and inclusions in the protective gas atmosphere. The quantitative control rules for process parameters include: increasing laser power and reducing scanning speed can weaken unfused porosity defects, while excessive power will exacerbate keyhole formation; adopting a single-layer remelting + multi-layer gradient overlapping strategy can significantly close residual pores between layers; strictly controlling the purity of the protective gas to ≥99.99% and the ambient humidity to <30%RH can prevent the formation of gas-included pores.

[0020] Step S4, Integrated Defect Closed-Loop Repair: Based on the mechanism models and control criteria of three typical defects—grain structure, hot cracks, and porosity—the complete set of laser additive cladding process parameters is optimized and matched to achieve defect-free and precise repair of the damaged area of ​​the turbine blade. The range of the complete set of optimized process parameters is as follows: laser power 800–1800W, scanning speed 5–15mm / s, overlap rate 30%–50%, powder feeding rate 8–20g / min, substrate preheating temperature 200–400℃, and single-layer remelting times 1–2 times.

[0021] This embodiment focuses on metal laser additive manufacturing repair and joining, such as centimeter-scale surface repair of shell parts and large-size forming joining of aluminum alloys using SLM (Silicon-Modulated Laser) technology. It finds that defect control and subsequent optimization of microstructure and properties are the core of performance control in thin-walled large-size additive manufacturing, and proposes an active defect control method based on temperature field regulation. In-depth research has been conducted on the transformation mechanism of microstructure and properties at the interface of heterogeneous materials, the gradient performance design method for interface microstructure control, and the refined characterization of defects. Notably, the Al-Cu-Mg-Si (2A50) alloy based on L-DED, after heat treatment, has achieved strength comparable to forgings (tensile strength > 330 MPa). Currently, challenges such as additive manufacturing of large integral structural beams, protective grilles for aero-engines, high-performance repair of engine casing parts, and full-size surface strengthening of nuclear power plant guide structures have been solved. Figure 2 , Figure 3 and Figure 4 As shown.

[0022] This invention conducts numerical simulation studies on the temperature and stress fields of additive manufacturing, such as... Figure 5 As shown, residual stress in the additive manufacturing deposited layer and substrate was measured using drilling and indentation methods, respectively. The influence of process parameters on residual stress and its distribution pattern were analyzed. Simulations were performed on the substrate and 3D printed components, and stress values ​​were obtained within the process window. Simultaneously, the scanning path was tested to obtain the residual stress distribution of the components. The research methods used above will be applied to the construction of a high-fidelity laser filament additive manufacturing simulation model in this project.

[0023] Example 2, a method for repairing defects in steam turbine blades using laser additive cladding, includes the following steps: Step 1: Analysis of Grain Structure Defect Mechanism and Orderly Growth Control. Focusing on martensitic stainless steel and heat-resistant steel commonly used in turbine blades, this step analyzes the critical conditions for the "columnar to equiaxed crystal" transformation and the grain orientation instability law, centering on the temperature gradient G and interface velocity R during laser melting deposition. It reveals the influence of heterogeneous nucleation, solute redistribution, and Marangoni convection on the microstructure morphology, thus clarifying the causes of grain skewness and orientation mixing. Then, for high-constraint, multi-pass, multi-layer deposition scenarios, combining thermo-elastic-plastic-phase transformation coupling analysis and crack morphology characterization, the main controlling factors of solidification cracks, liquefaction cracks, and reheat cracks are identified. The effects of dilution rate, thermal cycling superposition, and weak zone distribution on stress concentration and the conditions for continuous liquid film backfilling are clarified, and criteria for hot cracking tendency and suppression paths are proposed. Finally, by using in-situ high-speed imaging and infrared / optical monitoring, keyhole-type and non-fused-type pores are distinguished, and the effects of power, scanning speed, remelting strategy, overlap rate, powder / wire feeding rate, and gas protection / humidity on pore formation and disappearance are quantified, so as to construct a pore formation mechanism and control criteria for engineering applications.

[0024] For martensitic stainless steel and heat-resistant steel substrates for turbine blades, the temperature gradient G and solid-liquid interface velocity R of the molten pool during laser melting deposition were collected to establish a critical mathematical model for the columnar-equiaxed crystal transformation coupled with GR, and the grain orientation instability law was analyzed. Combining heterogeneous nucleation kinetics, solute element redistribution mechanism and molten pool Marangoni convection effect, the causes of grain skewness and mixed orientation were elucidated. By matching the GR parameter threshold, the ordered growth of grain structure in the deposited layer was achieved. The critical mathematical model for the GR-coupled columnar-equiaxed crystal transformation uses the grain growth undercooling as the core variable to limit the critical G / R ratio range for the transformation from columnar to equiaxed crystals. When G / R is higher than the upper critical value, the molten pool is dominated by directional columnar crystal growth; when G / R is lower than the lower critical value, a large number of heterogeneous nuclei are triggered, and uniform precipitation of equiaxed crystals is achieved. The grain orientation instability is specifically described as follows: During the rapid solidification process of laser deposition, Marangoni convection in the molten pool causes disordered melt flow, which disrupts the stability of the solid-liquid interface. At the same time, solute element segregation causes local nucleation barrier differences, ultimately leading to skewed grain growth direction and disordered mixed grain orientation. By controlling the molten pool convection intensity and stabilizing the interface temperature gradient, grain orientation defects can be eliminated.

[0025] Step Two: Research on the Correlation between Melt Pool Dynamic Characteristics and Process Parameters and Solidification Structure. First, based on finite element method / CFD and infrared thermography / pyrometer, a quantitative relationship was established between the molten pool temperature field characteristics (G, R, G / R, GR) and microstructure coarseness, orientation stability, and defect sensitivity, enabling visualized judgment of grain growth and instability. Then, high-speed imaging and flow field inversion were used to analyze the effects of Marangoni convection, backflow eddies, and splash redeposition on dendrite growth, inclusion trapping, and porosity nucleation. CCD / CMOS coaxial imaging was used to obtain the mapping between molten pool geometry (width / depth / length, wetting angle, stability) and microstructure / defects. Subsequently, through multi-factor orthogonal / response surface experiments, the influence of power, spot size, scanning speed, path strategy, overlap ratio, cladding head tilt angle, powder / wire feeding, shielding gas purity and flow field, preheating and interpass temperature, matrix microstructure state, and alloy composition on compaction forming, HAZ softening control, and crack / porosity suppression was summarized. Finally, by integrating multi-source data such as process parameters, molten pool thermal-optical features, microstructure and defect annotations, and mechanical properties, a multi-scale deep network with embedded physical priors is constructed. Robust feature extraction and domain adaptation are achieved by utilizing self-sparse coding and contrastive learning, and online predictions of microstructure and defect probabilities and uncertainty assessments are output.

[0026] Identification and Suppression of Multiple Types of Hot Cracks: For high-constraint, multi-layer laser deposition repair scenarios involving blades, a thermo-elastic-plastic phase transition multi-field coupled simulation model is constructed. Combined with crack micromorphology characterization, three types of thermal defects are distinguished: solidification cracks, liquefaction cracks, and reheat cracks. The effects of dilution rate, multi-layer thermal cycling, and the distribution of material weak zones on stress concentration in the repair area are clarified. The conditions for continuous liquid film backfilling are defined, and quantitative criteria for the tendency of hot cracking in laser additive repair are established, leading to corresponding process suppression paths. The main controlling factors for the three types of hot cracks are as follows: solidification cracks are caused by the enrichment of low-melting-point solute eutectic phases at the end of deposition and the tensile stress caused by the contraction of the continuous liquid film between grains; liquefaction cracks are caused by the remelting of low-melting-point phases in the heat-affected zone of the substrate due to multi-layer thermal cycling, leading to cracking of the intergranular liquid film under constraint stress; and reheat cracks are caused by the precipitation of brittle phases at grain boundaries and the release of residual stress during high-temperature tempering after deposition, leading to grain boundary cracking. The quantitative criterion for thermal cracking tendency is constructed based on the coupling of multiple parameters such as dilution rate threshold, number of thermal cycles, width of vulnerable area, and peak interface stress. When the calculated thermal cracking index is lower than the safety threshold, it is determined that there is no risk of thermal crack initiation in the repair area. The suppression path includes four core control methods: substrate preheating and temperature control, layer-by-layer cooling and pressure holding, gradient dilution rate process matching, and multi-layer scanning trajectory optimization.

[0027] Step 3, Pore Defect Classification, Identification, and Precise Control: An integrated detection method combining in-situ high-speed imaging, infrared monitoring, and optical microscopy is employed to accurately distinguish between keyhole-type pores and non-fused pores. The influence of laser power, scanning speed, remelting strategy, scanning overlap rate, powder / wire feeding rate, protective gas purity, and environmental moisture content on pore initiation, growth, closure, and demise is quantified, constructing an engineering-based pore formation mechanism model and standardized control criteria. The pore defect classification and identification criteria are as follows: Keyhole-type pores are irregular, large-diameter, deep holes caused by excessive laser energy leading to keyhole collapse of the molten pool and insufficient melt backfilling; non-fused pores are small, dispersed micropores caused by insufficient melting of powder / wire, insufficient interlayer overlap, and inclusions in the protective gas atmosphere. The quantitative control rules for process parameters include: increasing laser power and reducing scanning speed can weaken unfused porosity defects, while excessive power will exacerbate keyhole formation; adopting a single-layer remelting + multi-layer gradient overlapping strategy can significantly close residual pores between layers; strictly controlling the purity of the protective gas to ≥99.99% and the ambient humidity to <30%RH can prevent the formation of gas-included pores.

[0028] Step Four, Integrated Defect Closed-Loop Repair: First, addressing the high-temperature, dusty, and obstructed conditions of turbine blade repair, a multi-physical quantity dynamic signal acquisition system for the molten pool is constructed, integrating CCD / CMOS coaxial vision, high-speed cameras, infrared thermal imaging / pyrometers, and spectral monitoring. This system achieves time synchronization, calibration, and feature standardization, forming a process data lake. Then, relying on correlation models and multi-objective optimization (geometric accuracy, dilution rate, HAZ hardness band, compactness, low cracking), the conflicting relationships between "dense forming—low defects—uniform microstructure" are decoupled. A Bayesian optimization / evolutionary algorithm and physical constraint optimization are combined to obtain robust process windows and tolerance boundaries. Next, multimodal online features are input into the model to predict defect types and development trends in real time. Compensable thresholds are set and early warnings are triggered, linking power, powder / wire feeding, scanning speed, and path adaptive fine-tuning to construct a feedforward-feedback integrated closed-loop control. Finally, single / compound defect simulation tests on a bench are used to refine the samples and criteria, and actual blade installations are conducted in the industrial field for verification. The closed-loop calibration model and process window are then evaluated for consistency and robustness.

[0029] This application's embodiments are guided by the research on "characterization and monitoring of microscale defects driven by multi-physical quantity sensing of the molten pool in the laser additive manufacturing process of steam turbine blades". Addressing the engineering application needs of additive / cladding repair of steam turbine blades, it focuses on the causes and evolution of microscale defects such as microstructure orientation instability, hot cracks, and porosity. Firstly, it relies on multimodal online sensing such as infrared / pyrometers, visible light / high-speed imaging, and spectroscopy to acquire key features such as the temperature field, flow field, and geometric morphology of the molten pool. Then, it constructs an interpretable cross-scale correlation mechanism of "process-molten pool-microstructure / defect" and quantifies the characterization criteria. Then, by integrating physical priors and deep learning, a real-time characterization and prediction early warning model for microscale defects is established. By linking power, powder / wire feeding, and scanning strategies, a closed-loop quality control integrating feedforward and feedback is achieved, forming a robust and transferable repair process window and tolerance boundary. Furthermore, through full-link verification and data iteration on bench and in industrial fields, process data standards and replicable application solutions are accumulated. Ultimately, high consistency and low defects in the repair process, significant improvement in in-service reliability, and effective cost reduction are achieved, promoting the large-scale engineering implementation of laser additive repair for turbine blades.

[0030] Example 3: A method for repairing defects in 30Cr13 martensitic stainless steel steam turbine blades using laser additive cladding. This embodiment addresses the erosion defects of 30Cr13 martensitic stainless steel moving blades in thermal power turbines, and employs the laser additive cladding repair method of this invention. The specific steps are as follows: Step 1, analysis of grain structure defect mechanism and orderly growth control: real-time monitoring of molten pool temperature gradient G and interface velocity R, control of G / R ratio in the critical range of 250-350℃·s / mm, regulation of molten pool Marangoni convection intensity, optimization of solute redistribution uniformity, suppression of grain skew and orientation mixing, achieving orderly transition between columnar crystals and equiaxed crystals in the deposited layer, and improving the uniformity of structure by more than 40%; Step 2, Identification and Suppression of Multiple Types of Hot Crack Mechanisms: Construct a thermo-elastic-plastic phase transformation coupled model to identify the main defects in the repair process of this material as solidification cracks and liquefaction cracks, control the dilution rate at 8% to 12%, preheat the substrate to 300℃, adopt a layer-by-layer intermittent cooling process, calculate the hot cracking index to stabilize within the safe threshold, and completely eliminate the three types of hot crack defects. Step 3, Pore Defect Classification, Identification and Precise Control: Differentiate the types of pores in the repair process, optimize laser power to 1200W, scanning speed to 8mm / s, overlap rate to 40%, single-layer remelting once, control the purity of protective gas to 99.99% and the ambient humidity to 25%RH, completely eliminate keyhole-type pores, and control the porosity of unfused pores to below 0.5%. Step 4, Integrated Defect Closed-Loop Repair: The damaged area of ​​the blade is repaired using optimized complete process parameters. After metallographic and non-destructive testing, the repair layer is free of grain disorder, cracks, and excessive porosity. The interface bonding is good, and the mechanical properties match the substrate by more than 98%.

[0031] Example 4: Defect Repair of 12Cr1MoV Heat-Resistant Steel Steam Turbine Blades This embodiment addresses the high-temperature fatigue damage repair of 12Cr1MoV heat-resistant steel stator blades in nuclear power turbines, employing the control method of this invention: Step 1: Based on the GR critical model, adapt to the solidification characteristics of heat-resistant steel, adjust the G / R critical range of 180~280℃·s / mm, and eliminate the defects of coarse grains and disordered orientation in the high-temperature deposition process. Step 2: In view of the characteristic that the multi-layer deposition of this material is prone to reheat cracking, optimize the thermal cycling superposition process, avoid the high-temperature vulnerable area of ​​the material, establish targeted hot cracking criteria, and suppress the initiation of reheat cracks through gradient heat treatment process. Step 3: To address the issue of gas inclusions and pores in heat-resistant steel, the powder feeding rate was optimized to 15g / min, the scanning trajectory and remelting strategy were matched, and the humidity and protective gas parameters were quantified to achieve complete control over pore defects. Step four: After repair, the blades undergo high-temperature mechanical performance testing and long-term service simulation tests. The creep resistance and fatigue life are improved by more than 35% compared with traditional repair processes.

[0032] Preferred process parameters for the present invention are as follows: laser power 1000-1500W, scanning speed 6-12mm / s, overlap rate 35%-45%, powder feeding rate 10-18g / min, substrate preheating 250-350℃, single-layer remelting once, protective gas purity ≥99.99%, and ambient humidity 20%-30%RH. This parameter range is suitable for defect-free repair of most turbine blades made of martensitic stainless steel and heat-resistant steel, with the best defect suppression effect and the highest forming stability.

[0033] Example 5, please refer to Figure 6The control system for laser additive cladding repair of turbine blade defects includes: a molten pool parameter monitoring module 1, a microstructure characterization module 2, a multi-field coupled simulation module 3, a defect intelligent identification module 4, a process parameter control module 5, and a laser additive execution module 6. The molten pool parameter monitoring module 1 collects molten pool temperature, morphology, and environmental parameters in real time and transmits the data to the defect intelligent identification module 4 for online real-time defect detection. The detected data is then transmitted to the multi-field coupled simulation module 3 to construct a dynamic temperature and stress field model. The metallographic defect data detected offline by the microstructure characterization module 2 is synchronized to the multi-field coupled simulation module 3 and the defect intelligent identification module 4 to optimize the defect mechanism mathematical model. The multi-field coupled simulation module 3 calculates the grain transformation critical range, hot cracking tendency index, and porosity evolution law based on the real-time acquired data. The defect intelligent identification module 4 simultaneously outputs the real-time defect type and location. Simulation results and defect warning information are transmitted via OPC. The UA protocol transmits data to the process parameter control module 5. The process parameter control module 5 automatically optimizes the entire set of process parameters according to the defect control criteria, and sends control commands via the Profinet bus to drive the laser additive manufacturing execution module 6 to adjust the laser, motion, powder feeding, and gas path systems. The laser additive manufacturing execution module 6 transmits the actual operating status of the equipment back to the process parameter control module 5 in real time, completing the real-time closed-loop communication of the entire system's data acquisition, simulation identification, parameter control, and equipment execution. The hardware hierarchy of the control system for laser additive cladding repair of defects in turbine blades is divided into a bottom-level field acquisition layer, an intermediate processing layer, and a top-level execution control layer. The modules adopt a layered heterogeneous communication network, and the physical hardware connection, combined with industrial standardized communication protocols, completes real-time data interaction, command issuance, and simulation feedback closed-loop.

[0034] The molten pool parameter monitoring module 1 is used to collect molten pool temperature gradient, interface growth rate, and molten pool convection state parameters in real time; the defect intelligent identification module 4 can automatically distinguish grain defects, hot cracks, and two types of porosity defects, and match the corresponding process control scheme.

[0035] The molten pool parameter monitoring module 1 uses a high-speed infrared thermometer, a high-speed vision camera, a laser displacement sensor, and a gas temperature and humidity sensor to collect detection data. The sensor signal line is connected to the industrial edge gateway via shielded twisted pair cable. The gateway is equipped with three physical interfaces: RS485, Ethernet, and high-speed coaxial image cable. One coaxial video cable of the molten pool parameter monitoring module 1 is connected to the defect intelligent identification module 4, one RS485 bus is connected to the process parameter control module 5, and one gigabit network cable is connected to the multi-field coupling simulation module 3.

[0036] The microstructure characterization module has two offline auxiliary acquisition layers, including a metallurgical microscope, a scanning electron microscope, and a micropore flaw detector. The device is equipped with a gigabit Ethernet port and connects to the multi-field coupling simulation module 3 and the intelligent defect identification module 4 via a wired LAN switch. Offline detection data is transferred locally to a hard drive via USB 3.0 for data exchange. The multi-field coupling simulation module 3 serves as the central computing layer, an industrial control host, physically connected with dual gigabit network cards. One connection is to the melt pool parameter monitoring module 1 at the bottom edge gateway, and the other is to the microstructure characterization module 2 and the intelligent defect identification module 4 via the LAN switch. It also bidirectionally connects to the process parameter control module 5 via an Ethernet TCP / IP bus. The intelligent defect identification module 4 uses a GPU image processing workstation, directly connected to the high-speed camera in the melt pool via a high-speed coaxial cable, and connected to the multi-field coupling simulation module 3 and the process parameter control module 5 via gigabit Ethernet cables. A defect image sample library is stored on a local solid-state drive.

[0037] The process parameter control module 5 is the main control unit, employing a PLC industrial controller. It connects to the multi-field coupling simulation module 3 and the intelligent defect identification module 4 via Ethernet. An RS485 bus receives real-time temperature and flow rate data from the molten pool. A Profinet industrial bus hardwired connection directly connects to the laser additive manufacturing execution module 6, outputting analog / switching control signals. The laser additive manufacturing execution module 6 includes a fiber laser, a powder / fiber feeding mechanism, a six-axis motion platform, a preheating temperature control device, and a protective gas pressure regulating valve. The entire equipment drive unit is directly connected to the process parameter control module 5 PLC via a Profinet shielded cable, receiving motion and energy control commands such as power, scanning speed, overlap rate, and powder feeding rate.

[0038] The communication medium between the molten pool parameter monitoring module 1 and the defect intelligent identification module 4 is a high-speed coaxial video cable and a gigabit Ethernet. The communication protocol is the CoaXPress high-speed image transmission protocol and the UDP real-time data stream protocol. The real-time molten pool stream video, infrared temperature gradient G, solid-liquid interface velocity R, and molten pool Marangoni convection image are synchronously uploaded to the defect intelligent identification module 4. The defect intelligent identification module 4 sends image acquisition trigger and exposure adjustment control commands to the molten pool parameter monitoring module 1.

[0039] The communication medium between the molten pool parameter monitoring module 1 and the multi-field coupling simulation module 3 is gigabit industrial Ethernet, and the communication protocol is TCP / IP and OPC UA industrial IoT protocol. The original data of molten pool temperature field, flow field, ambient humidity and protective gas concentration are synchronized in real time and used as boundary input conditions for thermo-elastic-plastic-phase change coupling simulation. The simulation module periodically issues simulation correction parameters to calibrate the zero point of the monitoring sensor.

[0040] The communication medium between the microstructure characterization module 2, the multi-field coupled simulation module 3, and the defect intelligent identification module 4 is a gigabit local area network switch, and the communication protocols are FTP file transfer protocol and OPC UA; offline metallographic, crack, and pore microscopic images, grain size, crack length, and porosity detection reports are uploaded in batches; the multi-field coupled simulation module 3 calls the data from the microstructure characterization module 2 to correct the GR grain transformation model; the defect intelligent identification module 4 imports microscopic samples to complete the iterative training of the defect identification model.

[0041] The communication medium between the multi-field coupled simulation module 3 and the defect intelligent identification module 4 is Ethernet, and the communication protocol is bidirectional TCP Socket communication. The multi-field coupled simulation module 3 outputs the full-field stress distribution, thermal cracking index, and theoretical porosity evolution prediction results; the defect intelligent identification module 4 transmits back real-time online defect detection results, such as grain disorder, cracks, keyhole porosity location, and size; the data fusion of the two completes the defect risk classification prediction and generates a defect cause analysis dataset.

[0042] The communication medium for the multi-field coupling simulation module 3, the intelligent defect identification module 4, and the process parameter control module 5 is Industrial Ethernet, and the communication protocol is OPC UA protocol. The communication medium between the process parameter control module 5 and the laser additive manufacturing execution module 6 is the Profinet industrial dedicated shielded bus, and the communication protocol is the Profinet IO real-time industrial bus protocol.

[0043] This invention, for the first time, establishes a GR-coupled grain transformation critical model for martensitic stainless steel and heat-resistant steel materials in turbine blades, elucidating the influence mechanisms of heterogeneous nucleation, solute redistribution, and Marangoni convection on grain defects. It fundamentally reveals the causes of grain orientation disorder, filling the mechanistic gap in the precise control of grain structure during laser repair of turbine blades. Furthermore, it distinguishes the differentiated main control mechanisms of solidification, liquefaction, and reheating hot cracks, constructing a thermo-elastic-plastic phase transformation coupled analysis model and quantitative hot crack tendency criteria. This breaks away from the traditional empirical process control mode, enabling high-constraint multilayer deposition scenarios. Precise prediction and suppression of hot cracks; innovative differentiation between keyhole and non-fusion porosity defects; quantification of the influence of all core process parameters on porosity evolution; establishment of engineering control criteria; and significant reduction of porosity defect residue rate in the repair layer. The entire technology forms a closed-loop technical system from defect mechanism, quantitative criteria, process criteria to forming repair. The parameters are quantified and highly replicable, applicable to the repair of turbine blades of different damage degrees and specifications. It can significantly improve the blade repair qualification rate and service life, and meet the needs of industrialized mass remanufacturing production.

[0044] It should be noted that the embodiments described herein are only some embodiments of the present invention, and not all implementations of the present invention. These embodiments are merely illustrative and are intended only to provide a more intuitive and clear way to understand the content of the present invention, not to limit the technical solutions described herein. All other implementation methods that can be conceived by those skilled in the art without creative effort, as well as other simple substitutions and variations of the technical solutions of the present invention, without departing from the concept of the present invention, are within the protection scope of the present invention.

Claims

1. A method for repairing defects in steam turbine blades using laser additive cladding, characterized in that: The method for repairing defects in steam turbine blades using laser additive cladding includes the following steps: Step 1: Analysis of Grain Structure Defect Mechanism and Orderly Growth Control: For martensitic stainless steel and heat-resistant steel substrates for turbine blades, the temperature gradient G of the molten pool and the solid-liquid interface velocity R during the laser melting deposition process were collected. A critical mathematical model of columnar-equiaxed crystal transformation coupled with GR was established to analyze the grain orientation instability law. Combining heterogeneous nucleation kinetics, solute element redistribution mechanism and molten pool Marangoni convection effect, the formation causes of grain skewness and orientation mixing were clarified. By matching the GR parameter threshold, the ordered growth of grain structure in the deposition layer was achieved. Step 2: Identification and Suppression of Multiple Types of Hot Cracks: For the high-constraint, multi-pass, multi-layer laser deposition repair scenario of blades, a thermo-elastic-plastic-phase-transition multi-field coupled simulation model is constructed. Combined with the micro-morphological characterization of cracks, three types of thermal defects are distinguished: solidification cracks, liquefaction cracks, and reheat cracks. The effects of dilution rate, multi-layer thermal cycling superposition, and distribution of material weak areas on stress concentration in the repair area are clarified. The continuous liquid film backfilling molding conditions are defined, a quantitative criterion for the tendency of hot cracking in laser additive repair is established, and corresponding process suppression paths are formulated. Step 3, Porosity Defect Classification, Identification and Precise Control: An integrated detection method combining in-situ high-speed imaging, infrared monitoring and optical microscopy is adopted to accurately distinguish between keyhole-type pores and non-fused pore defects; the influence of laser power, scanning speed, remelting strategy, scanning overlap rate, powder feeding, wire feeding rate, protective gas purity and environmental humidity on pore initiation, growth, closure and extinction is quantified, and an engineering pore formation mechanism model and standardized control criteria are constructed; Step 4, Integrated Defect Closed-Loop Repair: Based on the mechanism models and control criteria of three typical defects—grain structure, hot cracks, and porosity—a complete set of laser additive cladding process parameters are matched to complete the defect-free and precise repair of the damaged area of ​​the turbine blade.

2. The method for repairing defects in turbine blades using laser additive cladding as described in claim 1, characterized in that: The critical mathematical model for the transformation from columnar to equiaxed crystals described in step one uses the grain growth undercooling as the core variable to limit the critical G / R ratio range for the transformation from columnar to equiaxed crystals. When G / R is higher than the upper critical value, the molten pool is dominated by directional columnar crystal growth; when G / R is lower than the lower critical value, a large number of heterogeneous nuclei are triggered, achieving uniform precipitation of equiaxed crystals.

3. The method for repairing defects in turbine blades using laser additive cladding as described in claim 2, characterized in that: The grain orientation instability described in step one is as follows: During the rapid solidification process of laser deposition, Marangoni convection in the molten pool causes disordered melt flow, which disrupts the stability of the solid-liquid interface. At the same time, solute element segregation causes local nucleation barrier differences, ultimately leading to skewed grain growth direction and disordered mixed grain orientation. By controlling the molten pool convection intensity and stabilizing the interface temperature gradient, grain orientation defects can be eliminated.

4. The method for repairing defects in turbine blades using laser additive cladding as described in claim 1, characterized in that: The three main controlling factors of hot cracks mentioned in step two are as follows: solidification cracks are caused by the enrichment of low-melting-point solute eutectic phases and the tensile stress of the continuous liquid film shrinkage between grains at the end of the deposition period; liquefaction cracks are caused by the remelting of low-melting-point phases in the heat-affected zone of the substrate due to multi-layer thermal cycling, and the cracking of the liquid film between grains under the action of restraint stress. Reheat cracks are caused by the precipitation of brittle phases at grain boundaries and the release of residual stress during the high-temperature tempering process after deposition, which triggers grain boundary cracking.

5. The method for repairing defects in turbine blades using laser additive cladding as described in claim 4, characterized in that: The quantitative criterion for thermal cracking tendency described in step two is constructed based on the coupling of multiple parameters, including dilution rate threshold, number of thermal cycles, width of the vulnerable zone, and peak interfacial stress. When the calculated thermal cracking index is lower than the safety threshold, it is determined that there is no risk of thermal crack initiation in the repair area. The suppression path includes four control methods: substrate preheating and temperature control, layer-by-layer cooling and pressure holding, gradient dilution rate process matching, and multi-layer scanning trajectory optimization.

6. The method for repairing defects in turbine blades using laser additive cladding as described in claim 1, characterized in that: The classification and identification criteria for porosity defects mentioned in step three are as follows: Keyhole type porosity is an irregular large-diameter deep hole, caused by excessive laser energy leading to the collapse of the keyhole of the molten pool and insufficient backfilling of the melt; Incomplete fusion type porosity is a fine and diffuse micropore, caused by insufficient melting of powder and wire, insufficient interlayer overlap, and inclusion of protective gas atmosphere.

7. The method for repairing defects in turbine blades using laser additive cladding as described in claim 6, characterized in that: The quantitative control rules for process parameters mentioned in step three include: increasing laser power and reducing scanning speed can weaken unfused porosity defects, while excessive power will exacerbate keyhole formation; adopting single-layer remelting and multi-layer gradient overlapping strategies can significantly close residual pores between layers; and controlling the purity of the protective gas to be greater than or equal to 99.99% and the ambient humidity to be less than 30%RH.

8. The method for repairing defects in turbine blades using laser additive cladding as described in claim 1, characterized in that: The complete set of optimized process parameters in step four are as follows: laser power 800~1800W, scanning speed 5~15mm / s, overlap rate 30%~50%, powder feeding rate 8~20g / min, substrate preheating temperature 200~400℃, and single-layer remelting times 1~2 times.

9. A control system for a method of repairing defects in turbine blades using laser additive cladding as described in any one of claims 1 to 8, characterized in that: The control system for laser additive cladding repair of defects in turbine blades includes: a molten pool parameter monitoring module, a microstructure characterization module, a multi-field coupled simulation module, a defect intelligent identification module, a process parameter control module, and a laser additive execution module. The molten pool parameter monitoring module collects molten pool temperature, morphology, and environmental parameters in real time and transmits the data to the defect intelligent identification module for online real-time defect detection. The detected data is then transmitted to the multi-field coupled simulation module to construct dynamic temperature and stress field models. Metallographic defect data detected offline by the microstructure characterization module is synchronized to the multi-field coupled simulation module and the defect intelligent identification module to optimize the mathematical model of the defect mechanism. The multi-field coupled simulation module calculates the critical range of grain transformation, the hot cracking tendency index, and the porosity evolution law based on the real-time acquired data. The defect intelligent identification module simultaneously outputs the real-time defect type and location. Simulation results and defect early warning information are transmitted via OPC. The UA protocol is transmitted to the process parameter control module; the process parameter control module automatically optimizes the entire set of process parameters according to the defect control criteria, and sends control commands through the Profinet bus to drive the laser additive manufacturing execution module to adjust the laser, motion, powder feeding, and gas path systems; the laser additive manufacturing execution module sends the actual operating status of the equipment back to the process parameter control module in real time.

10. The control system for repairing defects in turbine blades using laser additive cladding as described in claim 9, characterized in that: The molten pool parameter monitoring module is used to collect molten pool temperature gradient, interface growth rate, and molten pool convection state parameters in real time. The intelligent defect identification module can automatically distinguish between grain defects, hot cracks, and two types of porosity defects, and match the corresponding process control schemes.

Images