The transition from manual, labor-intensive processes to automated, precise, and flexible manufacturing systems represents a critical evolution for foundries aiming to enhance quality, efficiency, and competitiveness. In our facility, we primarily produce large-bore engine block casting parts. These components are characterized by extended machining cycles, necessitating the application of a primer coat to prevent corrosion during intermediate storage and handling. The inherent challenge lay in the vast diversity of our product portfolio—casting parts with significantly differing structures and dimensions. Our long-standing reliance on manual spray painting had become a substantial bottleneck, plagued by capacity limitations, high labor intensity, a demanding work environment, and increasing difficulties in recruiting skilled workers.
Consequently, the core problem we, as a team of process engineers, have focused on since the plant’s establishment is this: How can we develop and design an automatic spray painting process capable of adapting to a wide range of casting part specifications? The solution must enable precise robotic positioning and fully automated operation. Successfully addressing this is paramount for improving surface finish quality, enhancing product presentation, and strengthening our market position. This article details our systematic journey to achieve this goal by employing the IDDOV Six Sigma design methodology.
1. Introduction to the IDDOV Six Sigma Methodology
The IDDOV (Identify, Define, Develop, Optimize, Verify) methodology is a structured roadmap within the Design for Six Sigma (DFSS) toolkit. It is grounded in the principles of concurrent engineering and Design for X (DFX), focusing on the entire product lifecycle. The core objective of IDDOV is to systematically integrate critical customer requirements into the design of both the product and the process itself. This approach ensures rapid development, high quality, cost reduction, and effective problem-solving from the outset. This project is structured around its five distinct phases.
2. Phase I: Identify
The initial phase involved a rigorous assessment of the project’s necessity and feasibility. Increasing customer expectations for surface finish quality, coupled with internal strategic goals, made a compelling case for change.
2.1 Project Background and Definition
Customer demands (VOC) were clear: manual painting led to issues like missed spots, flash rust, and poor paint adhesion, resulting in 36 customer claims over three months. The business imperative (VOB) was to modernize operations, targeting an increase in automation from 4.37% to 12%. Furthermore, employee needs (VOE) highlighted the strenuous nature of the manual work and its inability to support planned capacity growth.
After evaluating alternatives like manual painting, robotic painting, and electrostatic powder coating, we selected robotic spray painting for its high automation potential, superior and consistent finish quality, and broad industry applicability. The project was formally chartered as “Automatic Spray Painting Line Process Development and Optimization.”
2.2 Project Scope and Planning
The project scope covered 30 distinct part types, with the largest casting part weighing up to 1150 kg and measuring 703 mm × 602.5 mm × 1083 mm. A cross-functional team was formed, including project leadership, process engineering, equipment specialists, and shop-floor operators. The project timeline was set from December 2021 to July 2022.
2.3 Risk Assessment and Mitigation
A proactive risk assessment was conducted using brainstorming techniques. We identified 13 potential risks, rating them based on impact and probability to prioritize mitigation actions.
| Risk Category | Impact | Probability | Risk Score | Mitigation Action |
|---|---|---|---|---|
| Epidemic Impact | 3 | 4 | 12 | Strictly adhere to epidemic policies; conduct meetings online. |
| Insufficient Personnel Skills | 3 | 3 | 9 | Implement specialized robot training for support and maintenance staff. |
| Inappropriate Process Flow | 3 | 4 | 12 | Conduct thorough analysis to define the optimal process flow. |
| Poor Fixture Consistency | 3 | 3 | 9 | Design robust fixtures and enforce strict acceptance criteria. |
3. Phase II: Define
This phase focused on translating the voice of the customer into measurable, actionable project goals.
3.1 Customer Needs Analysis
Through interviews and surveys, we collected 14 distinct customer needs (from internal and external stakeholders). These were analyzed using a Kano model to categorize them into Basic, Performance, and Excitement factors, providing insight into their impact on satisfaction.
3.2 Defining Critical-to-Quality Characteristics (Y’s)
The hierarchical analysis and Analytical Hierarchy Process (AHP) were used to weight the customer needs. The top priorities were distilled into two critical-to-quality (CTQ) outputs for the project.
| Customer Need | Measurement Standard | CTQ (Y) | Data Type | Goal |
|---|---|---|---|---|
| Excellent Product Quality / Consistent Batch Quality | Casting Painting Inspection Standard | Y1: Painting Quality Score | Continuous (Larger is better) | >96 points |
| On-Time Delivery | Daily Production Count | Y2: Maximum Daily Output | Discrete (Larger is better) | >50 units/day |
Baseline measurements confirmed Y1 at 81.28 points and Y2 at 20.235 units/day, indicating significant room for improvement.
4. Phase III: Develop
In this phase, we conceptualized and designed the detailed process and its parameters.
4.1 High-Level Process Flow Design
The macro-process flow was established as: Loading → Blow-Off → Spray Painting → Flash-Off → Drying → Cooling → Unloading. This flow is supported by core systems: Material Handling, Conveyance, Painting, Drying, and ancillary Energy, Control, and Environmental systems.
4.2 Translating Needs into Technical Parameters
The first House of Quality (HoQ) was constructed to map customer needs onto the key design systems. This confirmed the primary systems requiring detailed design: the Loading/Unloading system, Conveyance system, Painting system, and Drying system.
A second HoQ and Axiomatic Design analysis were then used to derive the importance and sequence of specific design parameters (X’s). This disciplined approach prevented design coupling and redundancy.
4.3 Detailed Design of Critical X’s
Detailed designs were developed for each critical parameter using brainstorming, benchmarking, TRIZ, and simulation.
X1 – Loading/Unloading Method: An automatic gantry system was designed for precise (within 2 mm) and safe handling of heavy casting parts.
X4 – Fixture (Skid) Design: This was a paramount challenge due to the need for minimal contact (to allow robot access to internal cavities) while ensuring stability and universality across 30 part types. Using TRIZ principles (particularly Spatial Separation), we evolved the design through several concepts.
The final innovative design utilized a connected-base skid with composite material pads on the top plane, allowing a single fixture family to securely locate casting parts of varying footprints.
X6, X7, X8, X9, X10 – Painting Parameters: A mechanistic analysis formed the foundation. The double-spray method was selected for uniformity. To prevent runs/sags, the relationship between film thickness and painting time was modeled. The probability of a defect is related to the single-pass film thickness (μ). A logistic regression model from test data was:
$$ \ln\left(\frac{P}{1-P}\right) = -10.1468 + 0.177322 \times \mu $$
Where \( P \) is the probability of a sag defect. Setting \( P=0.01 \) yields a maximum allowable single-pass thickness of ~31.3 μm. Key parameters were then refined via simulation:
$$ \text{Single-pass thickness: } \mu = \frac{q \cdot a \cdot (1-b) \cdot t}{v \cdot l \cdot \tan(\theta/2)} $$
$$ \text{Total painting time: } t = \frac{S \cdot c}{v \cdot l \cdot \tan(\theta/2)} $$
Where:
- \( q \): Fluid flow rate (set to 100 mL/min)
- \( a \): Paint solids content (constant)
- \( b \): Overspray loss factor
- \( v \): Robot gun speed (set to 0.2 m/s)
- \( l \): Stand-off distance (set to 200 mm, industry optimum)
- \( \theta \): Fan angle (set to 90°)
- \( S \): Total surface area
- \( c \): Number of passes (3 passes of double-spray)
The final design called for a paint viscosity of 20-35 s, achieving a target total film build of ~75 μm without sags.
X15 – Line Layout: A U-shaped layout with 24 stations over approximately 70 meters was designed to balance footprint and flow.
The complete set of design parameters was summarized, and a Design Failure Mode and Effects Analysis (DFMEA) was conducted to identify and address potential failure modes proactively.
| Parameter | Design Content / Target Value |
|---|---|
| X1: Loading Method | Automatic Gantry |
| X2: Conveyor Speed | 0.1 m/s |
| X4: Fixture Design | Universal Skid with Composite Pads |
| X6: Paint Path | Double-Spray, 3 Passes |
| X7: Paint Viscosity | 20-35 s |
| X8: Paint Flow Rate | 100 mL/min |
| X9: Nozzle Angle | 90° |
| X10: Gun Speed | 0.2 m/s |
| X11: Drying Time | ≥ 21.6 min |
| X12: Drying Temperature | 80 °C |
| X13: Cooling Time | ≥ 7 min |
| X14: Energy Source | Natural Gas Hot Air |
| X15: Line Layout | U-shaped, ~70m, 24 stations |
5. Phase IV: Optimize
With the line installed and commissioned, we conducted pilot runs to test and refine the design.
5.1 Initial Sample Testing and Paint Coverage Optimization
Initial samples showed excellent film quality on major surfaces, with superior uniformity compared to manual painting. However, deep recesses (cavities >40mm) showed insufficient coverage because the fixed stand-off distance (\( l \)) increased effectively in these areas, reducing film build according to the thickness formula.
Several solutions were tested:
- Reducing stand-off distance locally: Adjusted robot path; resulted in thin paint on cavity edges.
- Slowing gun speed (\( v \)) in recesses: Caused sags on adjacent surfaces.
- Increasing flow rate (\( q \)) dynamically: Led to poor atomization.
The optimal solution was a combined adjustment of \( l \) and \( \theta \)** within the recess path, slightly decreasing distance and adjusting the fan angle to focus the pattern. This successfully cured the coverage issue without creating defects.
5.2 Process and Error-Proofing Optimization
Cooling Safety: Validated that a 7-minute cool-down brought part temperature below 35°C, preventing burn hazards. An interlock was added to the PLC to prevent unloading before this time elapsed.
Part Identification: A vision system was implemented to compare the casting part on the skid against a pre-loaded image for the selected program. A mismatch triggers a line stop, preventing wrong-program painting.
5.3 Capacity Verification
Cycle times for each station were measured. The bottleneck was the Painting operation at 406 seconds. The theoretical daily capacity was calculated, confirming it met the Y2 target.
| Station | Quantity | Process Time (s) | Transfer Time (s) | Total Time (s) | Cycle Time (s) | Hourly Rate |
|---|---|---|---|---|---|---|
| Loading | 4 | 107 | 240 | 347 | 86.75 | 41.48 |
| Blow-Off | 1 | 60 | 180 | 240 | 240 | 15.00 |
| Spray Painting | 2 | 752 | 60 | 812 | 406.00 | 8.87 |
| Flash-Off | 2 | 90 | 120 | 210 | 105.00 | 34.29 |
| Drying | 5 | 1440 | 70 | 1510 | 302.00 | 11.92 |
| Cooling | 3 | 420 | 30 | 450 | 150.00 | 24.00 |
| Unloading | 1 | 57 | 0 | 57 | 57.00 | 63.16 |
| Total / First Part | – | 2926 | 700 | 3626 | – | – |
$$ \text{Max Daily Output} = \frac{\text{Daily Work Time} – \text{First Part Time}}{\text{Bottleneck Cycle Time}} + 1 = \frac{(8 \times 3600) – 3626}{406} + 1 \approx 54.1 \text{ units} $$
6. Phase V: Verify
A production run of 1600 casting parts was conducted to validate the optimized process.
Y1 – Painting Quality Score: The average score rose to 97.59 points, significantly surpassing the 96-point goal and the baseline of 81.28.
Y2 – Maximum Daily Output: A sustained output of 54.29 units/day was achieved, exceeding the 50 units/day challenge target.
All process documentation, including work instructions, fixture drawings, and parameter sheets, was standardized and deployed. Operator training was completed. The project delivered substantial quality and productivity gains, along with improved safety and ergonomics, achieving a positive financial return.
7. Conclusion
The application of the structured IDDOV methodology was instrumental in successfully developing and implementing a flexible robotic spray painting system for large, varied casting parts. We achieved a breakthrough in automation, eliminating a major manual bottleneck. The project significantly enhanced the consistency and quality of the primer finish on our casting parts, directly addressing customer demands. Furthermore, it alleviated heavy labor, improved the workplace environment, and provided a scalable solution for capacity growth. The technical learnings, particularly in flexible fixturing and adaptive robot pathing, have provided a valuable template for automating other processes like grinding on large casting parts, paving the way for a more intelligent, efficient, and sustainable foundry operation.