The quest for enhancing the performance and longevity of engineering components, particularly those subjected to harsh service environments, drives continuous innovation in surface engineering. Among these components, grey iron castings hold a significant position, especially in automotive applications such as brake discs, due to their excellent castability, good damping capacity, and favorable thermal conductivity. However, a primary limitation of grey iron casting components is their relatively poor corrosion and wear resistance, which can lead to premature failure, safety concerns, and economic losses from frequent replacements. Traditional repair methods like electroplating are often inadequate, offering limited performance enhancement and posing environmental challenges. Laser cladding has emerged as a superior alternative, enabling the deposition of high-performance alloys onto substrate surfaces to restore or even improve their functional characteristics. This study focuses on leveraging a multi-objective Genetic Algorithm (GA) to optimize the laser cladding process for depositing 316L stainless steel alloy onto HT250 grey iron casting substrates. The goal is to systematically identify the optimal combination of process parameters—laser power, scanning speed, and powder feeding rate—that yields a clad layer with superior geometric integrity, excellent macroscopic morphology, and enhanced hardness, thereby providing a robust solution for the repair and remanufacturing of worn grey iron casting parts.
1. Introduction and Problem Statement
Grey iron casting, characterized by its graphite flake microstructure, is a workhorse material in heavy industry. Its widespread use in brake discs, engine blocks, and machine tool bases stems from a valuable combination of properties. Yet, the very graphite structure that provides good machinability and vibration damping also creates pathways for corrosion and initiates wear under sliding contact. In automotive brake discs, constant exposure to moisture, road salts, and cyclical thermal-mechanical stresses accelerates surface degradation through rusting and scoring. This not only compromises braking efficiency but also raises maintenance costs. While complete replacement is an option, it is resource-intensive. A sustainable engineering approach is to repair and enhance the functional surface. Laser cladding, a directed energy deposition process, offers a precise method to fuse a metallurgically bonded, corrosion- and wear-resistant layer onto the grey iron casting substrate. The selection of 316L stainless steel powder as the clad material is strategic, as it provides excellent corrosion resistance, good toughness, and moderate hardness, forming a protective barrier on the vulnerable grey iron casting surface.
The primary challenge in laser cladding lies in the complex interplay between process parameters and the final clad quality. Suboptimal parameters can lead to defects such as pores, cracks, high dilution (excessive melting of the substrate), poor bonding, or undesirable geometric shapes, all of which undermine the performance of the repaired grey iron casting component. Therefore, a systematic optimization methodology is crucial. This work employs a Genetic Algorithm, a robust evolutionary computation technique inspired by natural selection, to navigate this multi-parameter, multi-objective optimization space. The GA is tasked with finding parameter sets that simultaneously maximize cladding efficiency, minimize heat-affected zone (HAZ) size and surface roughness, and maintain dilution within a specified functional range. The subsequent experimental validation and analysis provide a comprehensive understanding of how these optimized parameters influence the macro- and micro-scale properties of the 316L layer on the grey iron casting.
2. Experimental Methodology and GA Optimization Framework
2.1 Materials and Preparation
The substrate material used was HT250 grey iron casting, a common grade with a nominal hardness of 30 HRC. The specimens were machined into blocks of dimensions 55 mm × 55 mm × 30 mm. Prior to cladding, the surfaces were meticulously cleaned to remove oils and oxides and inspected using ultrasonic testing to ensure the absence of subsurface defects like pores or cracks. The cladding material was gas-atomized 316L stainless steel powder. Both substrate and powder were acclimatized in a laboratory maintained at 22°C ± 1°C for 24 hours before experimentation to ensure consistency. The chemical compositions of both materials are detailed in Table 1.
| Material | C | Si | Ni | P | S | Mn | Cr | Mo | O | Fe |
|---|---|---|---|---|---|---|---|---|---|---|
| 316L Powder | 0.018 | 0.92 | 11.3 | – | – | – | 15.0 | 2.50 | 0.33 | Bal. |
| HT250 Grey Iron | 3.15 | 1.78 | – | 0.08 | 0.12 | 0.78 | – | – | – | Bal. |
2.2 Laser Cladding System and Fixed Parameters
The laser cladding system integrated a 3 kW fiber laser with a coaxial powder feeding nozzle, mounted on a robotic arm for precise motion control. The powder feeder was a synchronized disc-type system. To isolate the variables for optimization, several parameters were held constant throughout the experiments based on preliminary trials and system specifications:
- Laser spot size: 6 mm × 3 mm (rectangular).
- Shielding gas (N₂) pressure: 0.35 MPa.
- Carrier gas (N₂) flow rate: 495 L/h.
- Standoff distance: Kept constant for a focused beam on the grey iron casting surface.
- Substrate preheating: None applied.
2.3 Genetic Algorithm for Multi-Objective Optimization
The core of the parameter selection process was a GA designed to optimize three key variable parameters within defined bounds:
- Laser Power (P): 2000 W to 3000 W.
- Scanning Speed (V): 8 mm/s to 12 mm/s.
- Powder Feeding Rate (F): 0.2 g/s to 0.8 g/s.
The optimization aimed to satisfy multiple, often conflicting, objectives derived from desired clad layer characteristics on the grey iron casting:
- Maximize Cladding Efficiency (E): A higher deposition rate is economically favorable. This is proportional to the powder feed rate and inversely related to scanning speed. A simplified model can be: $$ E \propto \frac{F}{V} $$
- Minimize Heat-Affected Zone (HAZ): A smaller HAZ preserves the base metal properties of the grey iron casting and reduces residual stresses. HAZ is generally proportional to energy input: $$ HAZ \propto \frac{P}{V \cdot D} $$ where D is the beam diameter.
- Minimize Surface Roughness (Rz): A smoother as-cladded surface reduces post-processing. Roughness is a complex function of melt pool dynamics and solidification.
- Control Dilution (η): Defined as the ratio of substrate melt depth to total clad height, it must be kept within 5% to 25% to ensure good bonding without excessive dilution of clad alloy properties by the grey iron casting substrate. Dilution increases with higher energy input per unit length.
The multi-objective function for the GA can be formulated as:
$$
\text{Find } X = [P, V, F] \text{ to } \mathbf{minimize} \quad \mathbf{F}(X) = \begin{bmatrix}
-f_E(P, V, F) \\
f_{HAZ}(P, V, F) \\
f_{R_z}(P, V, F)
\end{bmatrix}
$$
Subject to: $$ 5\% \leq \eta(P, V, F) \leq 25\% $$
$$ P_{min} \leq P \leq P_{max}, \quad V_{min} \leq V \leq V_{max}, \quad F_{min} \leq F \leq F_{max} $$
Here, \( f_E, f_{HAZ}, f_{R_z} \) are objective functions derived from empirical or analytical models relating the parameters to efficiency, HAZ size, and roughness. The GA works by creating a population of candidate parameter sets (chromosomes), evaluating their fitness based on the multi-objective function, and iteratively applying genetic operators—selection, crossover, and mutation—to evolve toward optimal solutions. The key parameters for the GA itself are listed in Table 2.
| Parameter | Value |
|---|---|
| Population Size | 20 |
| Maximum Generations | 100 |
| Crossover Probability | 0.55 |
| Mutation Probability | 0.15 |
| Selection Method | Tournament Selection |
The algorithm flow is summarized as follows:
- Initialization: Randomly generate an initial population of parameter sets within bounds.
- Fitness Evaluation: For each set, calculate objectives (E, HAZ, Rz, η) using a pre-calibrated process model (e.g., Response Surface Model based on prior DoE).
- Pareto Ranking & Selection: Rank solutions based on non-domination (Pareto front). Select parents for reproduction favoring higher-ranked solutions.
- Crossover & Mutation: Create offspring by combining parent parameters (crossover) and randomly altering some values (mutation).
- New Population: Form a new generation from elite parents and offspring.
- Termination: Repeat steps 2-5 until the maximum generation count is reached. The final Pareto-optimal front represents the best compromise solutions.
The convergence of the GA is indicated by the stabilization of the average and best fitness values across generations. After optimization, the GA output a set of Pareto-optimal parameter combinations. One of these prominent solutions, along with variations around it to study individual parameter effects, was selected for experimental verification. The experimental matrix derived from the GA output is presented in Table 3.
| Test Group | Laser Power, P (W) | Scanning Speed, V (mm/s) | Powder Feed Rate, F (g/s) |
|---|---|---|---|
| 1 | 2000 | 8 | 0.25 |
| 2 | 2000 | 9 | 0.50 |
| 3 | 2000 | 10 | 0.75 |
| 4 | 2400 | 10 | 0.50 |
| 5 | 2400 | 9 | 0.25 |
| 6 | 2400 | 8 | 0.75 |
| 7 | 2800 | 10 | 0.25 |
| 8 | 2800 | 8 | 0.50 |
| 9 | 2800 | 9 | 0.75 |
2.4 Characterization Techniques
After laser cladding on the grey iron casting substrates, the specimens were subjected to comprehensive characterization:
- Macroscopic Morphology: Visual and stereo-microscopic inspection for surface smoothness, presence of cracks, pores, and overall bead shape consistency.
- Geometrical Dimensions: The clad bead width (W) and height (H) were measured using a digital vernier caliper. Each reported value is the average of at least five measurements along the clad track length.
- Hardness: Rockwell hardness (HRC) measurements were taken directly on the top surface of the clad layer using a standard hardness tester. A minimum of five indentations were made per sample.
- Dilution Assessment: Cross-sections of selected samples were prepared by standard metallographic techniques (mounting, grinding, polishing). The dilution was calculated from optical microscope images using image analysis software: $$ \eta = \frac{A_{substrate\ melt}}{A_{total\ clad}} \times 100\% $$ where A represents the cross-sectional area.
3. Results, Analysis, and Discussion
3.1 GA Optimization Results and Convergence
The Genetic Algorithm successfully converged to a set of optimal solutions. The progression of the average and best fitness values across generations demonstrated typical evolutionary behavior: rapid initial improvement followed by stabilization as the population approached the Pareto-optimal front. The stabilization indicated that further generations were unlikely to yield significantly better solutions. From the final Pareto front, a key solution was identified: Laser Power = 2800 W, Scanning Speed = 10 mm/s, Powder Feed Rate = 0.25 g/s (corresponding to Group 7 in Table 3). This parameter set promised an optimal balance of high hardness, good geometrical form, and controlled dilution for the grey iron casting repair application.
3.2 Analysis of Clad Layer Geometry on Grey Iron Casting
The geometry of the single-track clad bead is a direct indicator of process stability and energy distribution. The measurements of width (W) and height (H) for all nine parameter sets are summarized in Table 4 and analyzed below.
| Group | Width, W (mm) | Height, H (mm) | Aspect Ratio (W/H) |
|---|---|---|---|
| 1 | 3.70 ± 0.15 | 0.26 ± 0.02 | 14.23 |
| 2 | 3.15 ± 0.12 | 0.28 ± 0.02 | 11.25 |
| 3 | 2.50 ± 0.10 | 0.18 ± 0.02 | 13.89 |
| 4 | 4.10 ± 0.18 | 0.35 ± 0.03 | 11.71 |
| 5 | 4.45 ± 0.17 | 0.48 ± 0.03 | 9.27 |
| 6 | 4.90 ± 0.20 | 0.61 ± 0.04 | 8.03 |
| 7 | 4.80 ± 0.18 | 0.42 ± 0.03 | 11.43 |
| 8 | 5.30 ± 0.22 | 0.60 ± 0.04 | 8.83 |
| 9 | 5.10 ± 0.21 | 0.58 ± 0.04 | 8.79 |
Effect of Scanning Speed (V): The most pronounced influence on bead geometry was observed from the scanning speed. For a constant laser power and powder feed rate, an increase in scanning speed decreases the linear energy input $$ \left( \frac{P}{V} \right) $$ delivered to the grey iron casting substrate. This results in a smaller, shallower melt pool. Consequently, both the clad width and height decrease with increasing scanning speed. This is clearly seen in Groups 1-3 (P=2000W, F varying): as V increased from 8 to 10 mm/s, W decreased from 3.70 mm to 2.50 mm and H showed a decreasing trend. Conversely, a lower scanning speed (e.g., Group 6 and 8 at 8 mm/s) resulted in the largest widths and heights due to higher energy input per unit length.
Effect of Laser Power (P): Increasing laser power, while keeping V and F constant, increases the energy input, leading to a larger melt pool and greater melting of both the powder and the grey iron casting substrate. This typically increases the clad width and can also affect height, though the relationship with height is more complex as it also depends on powder capture efficiency. Comparing groups with similar V and F but different P (e.g., Group 3 vs Group 4 vs Group 7, where V=10 mm/s, F~0.25-0.5 g/s), the width increased significantly with power: 2.50 mm @2000W, 4.10 mm @2400W, 4.80 mm @2800W.
Effect of Powder Feed Rate (F): At a constant energy input, a higher powder feed rate supplies more material to the melt pool. If the energy is sufficient to melt all the added powder, the primary effect is an increase in clad height. However, if the powder feed is too high for the available energy, unmelted particles can occur, or the width may even decrease as energy is dissipated in heating more powder. In our results, for medium/high power levels (2400W, 2800W), higher feed rates (e.g., Group 6, F=0.75 g/s) generally led to greater heights compared to lower feed rates at similar speeds.
The aspect ratio (W/H) is an important metric for overlapping tracks in multi-layer cladding. A moderate aspect ratio (e.g., ~10-12) is often desirable to minimize inter-track porosity. The GA-optimized parameter set (Group 7) yielded an aspect ratio of 11.43, which is within a favorable range for further deposition if needed on the grey iron casting component.
3.3 Analysis of Surface Hardness (HRC)
The Rockwell hardness of the clad surface is a critical performance metric, indicating resistance to indentation and wear. The measured HRC values are presented in Table 5 and analyzed in the context of process parameters.
| Group | Hardness, HRC | Std. Dev. |
|---|---|---|
| 1 | 30.5 | ± 0.8 |
| 2 | 29.8 | ± 1.0 |
| 3 | 28.2 | ± 1.2 |
| 4 | 33.1 | ± 0.7 |
| 5 | 34.7 | ± 0.6 |
| 6 | 32.9 | ± 0.9 |
| 7 | 37.6 | ± 0.5 |
| 8 | 35.2 | ± 0.8 |
| 9 | 33.8 | ± 0.7 |
The hardness of the 316L clad layer is influenced by its microstructure, which in turn is governed by the solidification conditions determined by the process parameters. The key factors are cooling rate and dilution with the grey iron casting substrate.
Effect of Laser Power and Scanning Speed (Cooling Rate): The cooling rate $$ (\frac{dT}{dt}) $$ is inversely related to the energy input per unit length. Higher power and lower speed increase energy input, leading to a larger melt pool and a slower cooling rate, which can result in coarser microstructures and potentially lower hardness. Conversely, lower power and higher speed promote rapid solidification, yielding finer microstructures and higher hardness. However, this is counterbalanced by dilution.
Effect of Dilution (Compositional Change): Dilution with the high-carbon grey iron casting substrate is a crucial factor. The 316L alloy is a low-carbon austenitic stainless steel. Melting of the grey iron casting substrate introduces carbon and other elements into the clad melt pool. Upon rapid cooling, this can lead to the formation of hard martensitic and carbide phases within the austenitic matrix, significantly increasing hardness. Therefore, a certain level of dilution is beneficial for hardness, but excessive dilution can compromise corrosion resistance and toughness.
Interpretation of Results: Group 7 (P=2800W, V=10 mm/s, F=0.25 g/s) achieved the peak hardness of 37.6 HRC. This parameter set represents a strategic balance:
- The high laser power (2800W) and moderate-high scanning speed (10 mm/s) create a sufficient but not excessive energy input, leading to a reasonably fast cooling rate that refines the microstructure.
- The relatively low powder feed rate (0.25 g/s) means that for a given energy input, a higher proportion of the melt pool volume comes from the melted substrate compared to a high feed rate scenario. This results in a controlled, moderate level of dilution (estimated within the target 5-25% range), enriching the clad layer with carbon from the grey iron casting and promoting the formation of hardening phases without causing defects like cracking.
In contrast, at low laser power (2000W, Groups 1-3), the hardness is lower (~28-31 HRC). Here, the energy input may be insufficient to achieve adequate fusion and the desired level of dilution or may lead to incomplete melting of powder. At high powder feed rates with high power (e.g., Group 9), dilution decreases because more of the melt pool is composed of the fed 316L powder, leading to a hardness closer to that of wrought 316L but lower than the optimally diluted case for this grey iron casting system.
The hardness can be conceptually modeled as a function of cooling rate and dilution:
$$ HRC \approx f\left(\frac{P}{V}, \eta \right) = f_{cooling}\left(\frac{P}{V}\right) + f_{dilution}(\eta) $$
where \( f_{cooling} \) has a negative correlation with \(P/V\) (finer structure from faster cooling increases hardness) and \( f_{dilution} \) has a positive but saturating correlation with \(\eta\) up to an optimal point, after which properties degrade.
3.4 Macroscopic Morphology and Defect Analysis
Visual and microscopic inspection of the clad tracks revealed clear correlations with process parameters. The macroscopic quality is paramount for the grey iron casting repair application, as it minimizes post-processing.
- Low Power Groups (1-3, P=2000W): These tracks appeared less consistent, with visible spatter and a less smooth surface. The tracks were narrower and flatter. At the highest scanning speed and feed rate (Group 3), the track was discontinuous in some areas, indicating borderline energy for stable melting, which is undesirable for coating a grey iron casting surface.
- Medium Power Groups (4-6, P=2400W): A significant improvement in track appearance was observed. The tracks were continuous, with a more regular, convex shape. Surface smoothness improved, though some minor surface rippling was present.
- High Power GA-Optimized Group (7, P=2800W, V=10mm/s, F=0.25g/s): This parameter set produced the most desirable macroscopic morphology. The clad track was fully dense, continuous, and exhibited a smooth, shiny surface with a regular geometrical cross-section. No visible cracks, pores, or balling effects were observed, indicating excellent process stability and melt pool control on the grey iron casting substrate.
- Other High Power Groups (8 & 9): These also produced good, continuous tracks. However, at lower scanning speeds (Group 8), the track was very wide and thick, which might not be efficient for precision repair. With high powder feed (Group 9), the surface showed slightly more roughness compared to Group 7.
The superior morphology of Group 7 can be attributed to an optimal balance between energy input and mass input. The specific combination likely resulted in a stable melt pool with adequate fluidity to smooth out surface tension effects (leading to a shiny surface) but sufficiently rapid solidification to maintain shape definition without excessive spreading on the grey iron casting.
3.5 Synthesis: The Optimal Parameter Set for Grey Iron Casting
Integrating the analyses of geometry, hardness, and morphology, the GA-derived parameter set for Group 7 (Laser Power = 2800 W, Scanning Speed = 10 mm/s, Powder Feed Rate = 0.25 g/s) is conclusively identified as the optimal configuration for laser cladding 316L onto HT250 grey iron casting under the conditions of this study. This set achieves:
- Superior Hardness (37.6 HRC): A ~25% increase over the base grey iron casting hardness (30 HRC), indicating significantly improved wear resistance.
- Excellent Macroscopic Morphology: A dense, crack-free, smooth, and geometrically regular clad bead, reducing the need for post-cladding machining.
- Favorable Geometry: A clad bead with a width of ~4.8 mm and height of ~0.42 mm, providing adequate coverage and an aspect ratio suitable for multi-track overlap if required.
- Controlled Process: The parameters imply a controlled dilution within the target range, ensuring strong metallurgical bonding without overly compromising the clad alloy chemistry.
The success of this parameter set validates the effectiveness of the Genetic Algorithm as a tool for multi-objective optimization in complex processes like laser cladding on challenging substrates such as grey iron casting. The GA efficiently navigated the parameter space to find a solution that a purely trial-and-error approach might have missed.
4. Conclusions and Implications for Grey Iron Casting Repair
This comprehensive investigation demonstrates a successful methodology for optimizing the laser cladding process of 316L stainless steel onto grey iron casting substrates, specifically HT250, using a Genetic Algorithm. The key findings are:
- GA Efficacy: The Genetic Algorithm proved highly effective in solving the multi-objective optimization problem, converging to a Pareto-optimal front from which a superior process parameter set was extracted for experimental validation on grey iron casting.
- Parameter-Property Relationships: The experimental study elucidated clear relationships:
- Scanning Speed is the most dominant parameter affecting clad bead geometry (width and height) on the grey iron casting, with an inverse relationship.
- Surface Hardness is determined by a complex interplay between cooling rate (affected by P and V) and dilution with the carbon-rich grey iron casting substrate. An optimal combination of high power, moderate-high speed, and low powder feed rate was found to maximize hardness through controlled dilution and a refined microstructure.
- Macroscopic Quality requires sufficient energy density (P/V) to ensure stable melting and good fluidity, but not so much as to cause excessive spreading or evaporation. The GA-optimized set achieved this balance perfectly for the grey iron casting.
- Optimal Parameters: The identified optimal parameters—2800 W laser power, 10 mm/s scanning speed, and 0.25 g/s powder feed rate—produce a 316L clad layer on grey iron casting with exceptional properties: a high Rockwell hardness of 37.6 HRC, excellent surface morphology, and a geometrically sound bead profile with an estimated dilution within the functional 5-25% window.
Practical Implications: This work provides a validated, optimized laser cladding recipe specifically tailored for repairing and enhancing grey iron casting components, such as automotive brake discs. The significant improvement in surface hardness directly translates to better resistance against abrasive wear, while the 316L composition offers enhanced corrosion protection compared to the bare grey iron casting. The methodology—combining GA optimization with systematic experimental analysis—establishes a framework that can be adapted to other material combinations and cladding objectives, advancing the field of laser-based repair and additive manufacturing for cast iron components. The result is a promising, efficient, and high-performance solution for extending the service life of valuable grey iron casting parts, contributing to sustainability through remanufacturing.