1. Introduction
The global energy landscape is undergoing a profound transformation, shifting towards low-carbon, pollution-free, safe, affordable, and sustainable energy sources. Wind power has emerged as one of the fastest-growing renewable energy technologies, driven by the urgent need to reduce carbon dioxide emissions and combat climate change. As a result, the wind power casting industry has experienced remarkable growth, with global wind turbine casting production reaching approximately 4.8 million tons, accounting for about 4.5% of total global casting output in 2017. The Chinese market alone represents one-third of global wind power business, and the cumulative installed capacity of wind power in China has grown from 148 GW in 2016 to over 210 GW in 2018.
Wind turbine castings are critical components that connect the blades, support the nacelle, and bear complex and substantial alternating loads during operation. These components, including hubs, bases, bearing housings, and main shafts, must operate reliably for at least 20 years under harsh environmental conditions, demanding extremely high quality standards. However, the wind turbine casting industry faces significant challenges in quality control. Among various quality defects, slag inclusion has been identified as one of the most persistent and costly issues affecting wind turbine castings. This paper presents a comprehensive analysis of slag inclusion defects observed in V112 series products, including the base frame (BF), hub, and bearing housing (MBH), produced at my company, and details the systematic approach taken to identify root causes and implement effective solutions.
2. Equipment, Materials, and Process Overview
The production of wind turbine castings requires sophisticated equipment and carefully selected materials. The main production equipment used in our facility includes medium-frequency induction furnaces manufactured by Inductotherm (four 12-ton and two 6-ton units), 6-ton, 13-ton, and 25-ton pouring ladles from Finnish Technical SSK, overhead cranes from Konecranes, mixers from British Omega, and various auxiliary equipment such as molding conveyors, shakeout machines, and casting grinders.
Table 1 lists the principal raw materials used in wind turbine casting production:
| Material | Specification | Supplier |
|---|---|---|
| Pig iron | Carbon ≥ 4.0% | Benxi, Liaoning |
| Steel scrap | Carbon < 0.5% | Shanghai Shengda |
| Inoculant | Silicon 70-74% | Shanghai Elkem |
| Carburizer | 100% carbon | Wuxi Heli |
| Covering agent | 4L/barrel | Xuzhou Kaijierui |
| Nodulizer | 10-30mm particle size | Ningxia Elkem |
| Resin | Furan resin | Suzhou Foseco |
| Hardener | Vm100 & 200 | Jinan Shengquan |
| Coating | Baume degree ≥ 90% | Suzhou Foseco |
The wind turbine castings discussed in this paper are manufactured using the sand mold casting process, which is the most suitable method for producing large, complex-shaped components. The V112BF product, which is the primary focus of this study, weighs approximately 16.5 to 17 tons and requires extremely tight quality control. The production process flow involves several critical stages: pattern making, molding with furan resin sand, core assembly, metal melting, pouring, cooling, shakeout, grinding, and non-destructive testing.
The material specification for wind turbine castings is typically QT400-18AL grade ductile iron, which requires a minimum tensile strength of 400 MPa, minimum elongation of 18%, and must satisfy low-temperature impact toughness at -20°C. The large-section ductile iron castings, typically with wall thicknesses ranging from 100mm to 450mm, present unique challenges in terms of solidification control and defect prevention. The graphite morphology of thick-section ductile iron shows that graphite spheroidization must be carefully controlled to achieve the required mechanical properties and to minimize the risk of slag-related defects.
3. Current Production Status and Defect Analysis
As a professional manufacturer of high-quality large wind power castings, our company produces components ranging from 2 to 20 tons, including hubs, base frames, bearing housings, and main shafts. The production facility is organized into five interconnected production lines, with a sixth line dedicated to non-destructive testing. The production network is designed to follow a logical flow: raw materials enter through the melting area, molten iron is processed and transported to the molding area, castings are produced and undergo initial processing in the shakeout and fettling area, and finally, all products pass through comprehensive inspection before dispatch.
Statistical analysis of quality defects across the wind turbine casting product range reveals that slag inclusion consistently ranks among the top three defect types, along with sand inclusion and shrinkage. These three defect categories together account for over 70% of all quality failures observed in production. The slag inclusion defects manifest in various forms: sometimes appearing as dark, non-metallic particles on the casting surface; other times hidden within the casting internal structure, only to be discovered during subsequent machining operations.
The presence of slag inclusion defects in wind turbine castings, a form of sand foundry defect, not only compromises the aesthetic quality and dimensional accuracy but, more critically, severely undermines the mechanical integrity and fatigue life of the component. For critical load-bearing parts such as wheel hubs and main frames, such defects can lead to catastrophic failure during operation. Therefore, eliminating or significantly reducing slag inclusion defects is of paramount importance for ensuring both product reliability and the sustainable development of the wind turbine casting industry. The challenge is substantial because large-section ductile iron castings inherently have long solidification times, which intensifies the tendency for the formation of secondary oxidation slag. During the spheroidization treatment, the addition of magnesium and rare-earth elements creates favorable conditions for the development of these harmful non-metallic inclusions.
4. Root Cause Analysis of Slag Inclusion Defects
To systematically address the slag inclusion issues observed in V112 series products, our quality engineering team employed a multi-faceted analytical approach. This approach integrated several established quality tools and methodologies, including the 8-Discipline (8D) problem-solving method, High Potential Incident (HIPO) investigation, Fishbone (Ishikawa) diagrams, and the 5WHY analysis technique. These tools were used to analyze each stage of the internal casting process, from the incoming raw materials through melting, molding, pouring, and finishing operations.
Figure 1 shows an example of a sand foundry defect found on a casting surface.
4.1 Analysis of Melting and Spheroidization Process Changes
Our investigation revealed that the slag defect rate for V112BF products had dramatically increased from 1.1% in April 2018 to 18.8% by December 2018. This alarming upwards trend coincided with a process change aimed at increasing production throughput. To improve efficiency, the spheroidization treatment process was modified. In the original, established method, the molten iron was tapped into two separate 13-ton ladles before being combined into a 25-ton transfer ladle. The new, faster method involved tapping directly into the 25-ton ladle and significantly reduced the total time from the commencement of tapping to the start of pouring from approximately 31.7 minutes to 25.2 minutes. While this process change successfully shortened the treatment time by about 6-7 minutes, it inadvertently created favorable conditions for slag formation.
In the new method, the initial 9 tons of iron were tapped into the 25-ton ladle, where they reacted excessively with the nodulizing alloy in the bottom of the ladle. Then, the ladle was transported to the second furnace, where the second batch of 9 tons was tapped. By this point, the first batch of treatment alloy had already begun to oxidize, leading to a reaction that was less efficient and produced an increased quantity of slag particles. These slag particles, being heavier and having less time to float to the surface and coalesce, were more likely to be carried into the mold cavity during pouring, causing the observed defect.
4.2 Impact of Pouring Position and Basin Size
The analysis also implicated a change in the pouring position and gate basin configuration. The original process involved pouring from the long side of the casting, while the new process was implemented by rotating the entire setup by 90 degrees and pouring from the short side. This change was accompanied by a switch from a large to a significantly smaller pouring basin. Three-dimensional simulations using Magama software clearly demonstrated the problem. When pouring from the short side with the smaller basin, the operator had less control over the stream position. If the stream deviated by as little as 20 centimeters from the center of the sprue, the resulting vortex would draw slag particles floating on the surface of the molten iron in the basin directly into the gating system. These particles then traveled through the runner system and became embedded in the solidifying casting, resulting in internal slag inclusions. The simulation demonstrated that a larger, longer pouring basin provided a much more tolerant pouring zone, allowing the operator to maintain the melt stream on the sprue without creating conditions for slag entrainment.
4.3 Analysis of Pouring and Tapping Temperatures
To further reduce energy costs and shorten cycle times, a trial was conducted in which the tapping temperature was reduced from the original 1520°C to 1500°C. Of the five V112BF test castings produced with this reduced tapping temperature, three were found to have slag inclusion defects, a 60% defect rate. This was directly linked to the lower temperature, which led to a lower pouring temperature. At lower temperatures, the viscosity of the molten iron and, more importantly, the viscosity of slag particles increases, reducing their ability to float to the surface and be separated from the metal. Furthermore, lower temperatures can lead to premature solidification, trapping slag within the metal matrix.
4.4 Foundry Molding and Mold Compaction Problems
Physical inspection of defective castings and analysis of the internal non-conformance reports revealed a cluster of issues related to the molding process. These are all classic sources of the sand foundry defect. In several V112BF castings, ceramic tube fragments were found embedded in the casting surface. A detailed traceability investigation using 8D and HIPO methodologies, which included a review of process videos, found the root cause to be insufficient mold compaction, not a failure of the ceramic tube material itself. The compaction sequence performed by the operators was incorrect: they were manually compacting the sand around the base of the sprue tube after the main vibration compaction cycle had already finished. This late-stage compaction was ineffective and left the sand at the bottom of the sprue insufficiently dense. When the high-pressure molten iron entered the mold, it caused localized sand erosion and movement, which in turn exerted pressure on the base of the ceramic tube, leading to its fracture and subsequent entrainment in the casting.
Furthermore, a lack of proper mold compaction was identified in the region of the D-type stiffening ribs, located near the down sprue. The distance between the ribs and the sprue was less than 100mm, creating a narrow cavity that was extremely difficult to fill and compact effectively. The poor-quality mold in this area led to sand erosion, creating another form of the sand foundry defect. The investigation also found that the operator failed to secure the ceramic tube to the reinforcement bars within the mold with wire or cable ties, an essential step to hold the sprue in place during mold filling. A contributing factor was that the sand mold compactness was found to be non-uniform, which is known to be inversely related to the distribution of ferrostatic pressure during pouring. High ferrostatic pressure in areas with low mold compactness directly leads to mold wall movement, erosion, and the generation of slag. The requirement for more effective manual compaction was clearly identified as a critical need.
4.5 Management, Human Factors, and Missing Inspection
In addition to the technical and process problems, the investigation identified significant management and human factor issues. A key finding was that a defect had escaped detection and reached the customer due to a failure to follow inspection procedures. In one case, a casting was only visually inspected and was not subjected to the mandatory 100% ultrasonic testing. The defect was located in a geometric dead zone, such as a corner or edge, where the ultrasonic probe could not be placed due to its physical size and the inherent near-surface dead zone. This prevented the detection of a defect that was later uncovered during the customer’s machining process. Improvement actions were needed in our non-destructive testing procedure to ensure 100% coverage, and close coordination between the NDT operators and the fettling department was required.
Another major contributing factor was low employee morale. Between 2017 and 2018, the company implemented a large-scale capacity expansion project, from a design capacity of 40,000 tons to 80,000 tons per year. This led to changes in work patterns, work environment, and a period of instability in employee income due to the introduction of new standardized working procedures. This created dissatisfaction and resistance, which negatively impacted operational discipline and attention to detail, increasing the probability of human error.
5. Improvement Solutions and Design
Based on the comprehensive root cause analysis, a multi-pronged improvement plan was developed, encompassing process control, technical modifications, and management measures. The primary objective was to reduce the sand foundry defect rate for V112 products from the unacceptable level of 18.8% to below 5%.
5.1 Process Improvements for Melting and Spheroidization
The first step was to reverse the unsuccessful process change for the spheroidization treatment. A new method was developed and tested in which the first furnace’s 11 tons of molten iron at 1480°C was tapped into the 25-ton ladle, followed by immediately tapping a separate second furnace’s 7 tons of molten iron from the right side, also at 1480°C. This revised method generated an acceptable reaction, avoiding the over-treatment and slag formation observed in the previous process change. A key outcome of the trials was the establishment of a clear process window. The combined time for all the initial treatment steps was to be kept within 15 to 30 minutes. Tapping temperature was standardized at 1450-1490°C, and the pouring temperature was set at 1350-1400°C. Strict adherence to these parameters became a key control point in the updated Process Control Plan.
To ensure the stability of the process, a key area of focus was the standardization of raw material quality. It is essential to use pig iron with a carbon content ≥ 4.3%, and micro-elements must be tightly controlled: Silicon ≤ 0.8%, Manganese ≤ 0.5%, Phosphorus ≤ 0.04%, and Sulfur ≤ 0.02%. The steel scrap must have a carbon content of ≤ 0.5%, with gaseous elements oxygen ≤ 50 ppm, nitrogen ≤ 150 ppm, and hydrogen ≤ 4 ppm. These limits are critical to minimize the formation of oxide and nitride inclusions during melting and spheroidization, thereby directly reducing the potential for slag generation.
Controlling the addition of alloys was also important. For V112BF, the standard treatment for 18,600-18,700 kg of iron in the ladle requires:
| Product | Ladle treatment Iron (kg) | Nodulizer (kg) | Inoculant (kg) | Covering Agent (kg) | Antimony (g) |
|---|---|---|---|---|---|
| V112BF | 18,600-18,700 | 308 | 45 | 120 | 750 |
5.2 Modifications to Gating System and Pouring Equipment
The issue of the small pouring basin was solved by reverting to a larger, wider basin and pouring from the longer side. The larger basin provided a sufficient volume of molten iron to cover the sprue, which effectively prevented the vortex from reaching the surface slag layer and drawing it down into the gating system. The new pouring process is illustrated in Figure 2. This simple change proved highly effective in preventing surface slag from entering the casting mold.
Improvements were also made to the gating system design. The design was changed to increase the length of the pouring basin area, simplifying the operator’s ability to accurately maintain the molten metal stream within the sprue cup. Additionally, new, more robust designs were implemented for the gate and runner systems, incorporating multiple flow direction changes and additional reinforcement points. These design changes promoted smoother metal flow, minimized turbulence, and provided better mechanical integrity to resist the forces of the molten metal, all of which helped prevent erosion and the creation of sand foundry defects. The gating system calculation was performed using the standard formula for choke area calculation:
$$S_{内} = \frac{m}{\rho \tau \mu \sqrt{2g H_p}}$$
where \(S_{内}\) is the minimum cross-sectional area of the gate system, \(m\) is the mass of the casting, \(\rho\) is the density of molten iron (typically 7.0 g/cm³), \(\tau\) is the pouring time, \(\mu\) is the flow coefficient, \(g\) is the gravitational acceleration, and \(H_p\) is the static pressure head.
5.3 Molding Process Optimization
To address the molding defects, significant attention was paid to mandating and verifying correct mold compaction. The first action was to update and reinforce the Standard Operating Procedure (SOP) for molding. This new procedure explicitly instructs operators to manually compact the critical areas around the sprue tube base and all mold edges before, during, and after the main vibration compaction cycle. The manual compaction of the sprue area, in particular, must be performed using hand-held pneumatic rammers to achieve the required density, and this step must be performed before, not after, the vibration phase. The updated SOP was also modified to state that the D-type reinforcement ribs must be positioned at a minimum distance of 100 mm from the sprue to allow adequate space for proper compaction of the sand in that region.
Visual aids, including a laminated diagram of the correct compaction zones, were placed at each molding station. Additionally, the process now requires a photographic record to be taken of the compaction of the sprue area after it is completed. This provides visual evidence and helps guarantee the step is performed. The composition of the molding sand and the performance of the resin were also reviewed. The use of high-quality silica sand with a controlled particle size distribution (AFS 37.5-39.5) and the strict control of the furan resin and hardener mix ratios, taking into account the sand temperature, were identified as key parameters. The specification for reclaimed sand was defined with limits for dust and micro-content, and for micro-silica content.
| Indicator | Lower Limit | Upper Limit |
|---|---|---|
| Dust content (%) | – | 0.2 |
| Micro-content (%) | 0.063 | 0.18 |
| SiO₂ content (%) | 97 | – |
| Moisture content (%) | – | 1.0 |
We also established key parameters for inspecting the furan resin, including the control of free formaldehyde to below 0.3%, which is important for both quality and environmental safety. The coating selection and application process were also optimized. A key parameter was to control the coating’s Baume degree to approximately 60+2 for the pre-coat and to ensure a final coating thickness of 550-750 micrometers on the entire mold and core surface. This required careful monitoring and a robust training effort for the operators to apply the coating uniformly, paying special attention to corners, deep pockets, and other hard-to-reach areas.
5.4 Management, Inspection, and Human Factor Improvements
Recognizing the importance of the human element, a series of management actions was taken to boost employee morale and encourage participation. An employee suggestion program, called the Continuous Improvement (CI) program, was launched with a specific focus on quality issues. Employees were rewarded for identifying risks and proposing improvements, fostering a culture of ownership and proactive problem-solving. Team leaders and supervisors received communication and people management training to help them lead their teams more effectively through the change process.
To prevent the recurrence of missed defects, a comprehensive overhaul of the non-destructive testing (NDT) workflow was carried out. All products were now subject to 100% ultrasonic inspection, and a cross-checking system was implemented to ensure inspection integrity. The NDT technicians were encouraged to liaise more closely with the fettling and grinding team to better understand the outcome of their inspections and to help prioritize rework. A key part of this was the process of clearly marking the defect with a chalk outline and writing the dimensions and depth directly on the casting surface. This provides a clear visual guide for the grinding team. Furthermore, highly detailed checklists, known as “Firewall Inspection” sheet, were introduced to ensure that every critical step, from molding to final dispatch, was signed off and audited, guaranteeing that errors could be traced if found at any subsequent stage.
6. Results and Performance Evaluation
The implementation of the comprehensive improvement plan yielded significant and measurable results across multiple operational dimensions. The most direct and obvious outcome of tackling the sand foundry defect issue was a dramatic improvement in product quality. For the V112BF product line, the slag defect rate, which had peaked at 18.8%, dropped dramatically to an average of 4.2% within seven months of implementing the full suite of improvements. In the period from May 2019 to June 2019, quality tracking tables recorded zero instances of sand or slag inclusion defects in a sample of V112BF castings, demonstrating the effectiveness of the new procedures. The table below provides a comparison of the V112BF defect tracking before and after the main improvements.
| Defect Location | Sand/Slag Inclusion (Before) | Sand/Slag Inclusion (After) |
|---|---|---|
| Core internal cavity | 1 | 0 |
| Upper mold rib plate | 0 | 0 |
| Upper mold rib side | 0 | 0 |
| Upper mold flat surface | 0 | 0 |
| Upper mold ribs | 0 | 0 |
| Lower mold flat surface | 0 | 0 |
| Lower mold rib plate | 0 | 0 |
| Lower mold ribs | 0 | 0 |
This reduction in defects, directly attributable to overcoming a major source of the sand foundry defect, generated substantial cost savings. By avoiding scrap, the need for costly rework, and the potential for customer returns, the initiative avoided a potential loss of between 300,000 to 1.2 million RMB for a 60-piece production run. The improved quality also led to enhanced customer satisfaction and trust, which is crucial in the competitive wind power market. This quality improvement directly contributed to higher productivity per employee, which was a key factor in achieving the company’s capacity expansion targets, as shown in Table 5.
| Year | Productivity (kg/hour) |
|---|---|
| 2017 | 55 |
| 2018 | 65 |
| 2019 | 80 |
| 2020 | 95 |
The improvement in the management culture was also notable. The introduction of the CI program and the revised incentive plan encouraged a more engaged and motivated workforce. The continuous feedback and improved communication channels helped to identify emerging risks early and facilitated better cross-departmental cooperation. The success of the standardized NDT procedures, which included clear marking and traceability, enhanced the overall level of process control and quality assurance. The financial impact of these improvements was clear: the company achieved its cost reduction targets of 10 million RMB in both 2018 and 2019, which was in part due to the quality improvements and the reduction in waste.
| Year | Cost Savings (RMB) | Status |
|---|---|---|
| 2018 | 10,000,000 | Achieved |
| 2019 | 10,000,000 | Achieved |
| 2020 | 10,000,000 | Planned |
7. Conclusions and Future Prospects
Through a systematic and data-driven approach, this study successfully identified and addressed the root causes of slag inclusion defects in V112 series wind turbine castings, a major contributor to the overall sand foundry defect rate. The key conclusions are summarized as follows:
(1) The primary causes of the sand foundry defect were multi-factorial, including poor control of the melting and spheroidization process, especially the sequence and timing of ladle additions; excessively low tapping and pouring temperatures; a sub-optimal gating system design and pouring basin configuration; insufficient mold compaction due to a lack of standardized procedures and operator training; and inadequate inspection coverage that allowed defects to escape detection.
(2) The comprehensive improvement plan implemented, which involved the re-optimization of process parameters, simplification and standardization of work instructions, and a significant investment in employee engagement and management improvements, was highly effective. The slag inclusion defect rate for V112 products was reduced from a peak of 18.8% to 4.2% within seven months. This success demonstrated the power of a holistic problem-solving approach, combining technical expertise with a focus on human factors and robust management systems. The methods and solutions developed for V112 products have proven transferable and are now being applied successfully across the broader product portfolio.
(3) The work detailed in this paper confirms that controlling and eliminating slag and sand foundry defects requires a systematic, end-to-end approach. It is essential to control every stage, from the selection and inspection of raw materials, through precise control of melting and pouring parameters, to the discipline of the molding and inspection processes, and finally to the management systems and culture that support them. The success of this project highlights the importance of collaboration between engineering, operations, and quality assurance teams to achieve lasting improvement. Future work will continue to focus on further optimizing the process and extending these best practices across all product lines within the company, with a focus on continuous improvement.