In the field of modern manufacturing, nodular cast iron stands as a critical material due to its superior mechanical properties, including high strength, toughness, and wear resistance. However, the production of high-quality nodular cast iron is often hampered by casting defects such as shrinkage porosity, cavities, and poor nodularization, leading to increased scrap rates. To address these challenges, we have embarked on a comprehensive study leveraging thermal analysis technology to predict the microstructure and performance of nodular cast iron. This research focuses on establishing correlations between thermal analysis curve characteristics and key quality indicators, namely shrinkage tendency and nodularization effectiveness. By developing mathematical models through rigorous experimentation and statistical analysis, we aim to provide a reliable method for real-time quality prediction in industrial settings.
Thermal analysis involves monitoring the temperature changes during the solidification of molten iron, generating curves that reflect the kinetics of phase transformations. For nodular cast iron, these curves capture critical events such as the nucleation and growth of graphite spheroids and austenite. The characteristic values extracted from these curves, including temperatures and time ratios, serve as indicators of the solidification behavior. In this work, we conduct univariate experiments to investigate the effects of different nodularizers and inoculants on these characteristics. The goal is to elucidate how variations in additives influence the final microstructure and defect formation in nodular cast iron. Through this approach, we seek to enhance the controllability and consistency of nodular cast iron production, ultimately reducing waste and improving product reliability.
The significance of this study lies in its potential to transform quality assurance practices for nodular cast iron. Traditional methods for assessing nodularization and shrinkage are often offline, time-consuming, and subjective. By contrast, thermal analysis offers a rapid, in-line technique that can provide immediate feedback. Our research builds upon prior studies by systematically examining multiple factors and their interactions, rather than focusing on isolated variables. We employ advanced statistical tools, such as regression analysis via SPSS, to derive predictive models that link thermal analysis data to quantitative measures of quality. This holistic approach ensures that our findings are both scientifically robust and practically applicable, paving the way for smarter foundry operations.
Fundamentals of Thermal Analysis in Nodular Cast Iron
Thermal analysis curves are generated by recording the temperature of molten nodular cast iron as it cools and solidifies. These curves, along with their first derivatives, reveal distinct thermal events associated with phase transformations. For nodular cast iron, the solidification process typically involves the precipitation of primary phases followed by eutectic reactions where graphite spheroids form within an austenite matrix. The key characteristic values from these curves are defined as follows:
| Symbol | Description | Physical Meaning |
|---|---|---|
| \(T_{liq}\) | Liquidus temperature | Temperature at which solidification begins, corresponding to the first maximum in the second derivative. |
| \(T_{AL}\) | Austenite precipitation temperature | Inflection point on the cooling curve where the first derivative approaches zero. |
| \(T_{SEF}\) | Eutectic nucleation start temperature | Minimum of the first derivative between \(T_{AL}\) and \(T_{EU}\). |
| \(T_{EU}\) | Eutectic minimum temperature | Lowest temperature before recalescence, where the first derivative equals zero. |
| \(T_{ER}\) | Eutectic maximum temperature | Highest temperature after recalescence, where the first derivative equals zero. |
| \(T_{EM}\) | Maximum recalescence rate temperature | Temperature at which the recalescence rate is maximal, corresponding to the first derivative maximum. |
| \(\Delta T\) | Eutectic recalescence temperature | \(\Delta T = T_{ER} – T_{EU}\), indicating the extent of recalescence during eutectic reaction. |
| \(T_{ES}\) | Eutectic end temperature | Temperature at which solidification completes, marked by a minimum in the first derivative. |
| \(G_1/G\) | Early shrinkage time ratio | Ratio of time from \(T_{liq}\) to \(T_{EU}\) to total solidification time \(G\). |
| \(G_2/G\) | Graphitization expansion time ratio | Ratio of time from \(T_{EU}\) to \(T_{ER}\) to total solidification time \(G\). |
| \(G_3/G\) | Late shrinkage time ratio | Ratio of time from \(T_{ER}\) to \(T_{ES}\) to total solidification time \(G\). |
These parameters are crucial because they encapsulate the dynamics of solidification. For instance, \(\Delta T\) reflects the intensity of the eutectic reaction, which is linked to graphite formation. Similarly, time ratios like \(G_1/G\) and \(G_2/G\) relate to the sequence of shrinkage and expansion events that influence defect formation. In nodular cast iron, the balance between graphite expansion and metal contraction determines the likelihood of shrinkage porosity and cavities. By analyzing these characteristics, we can infer the internal quality of the cast material without destructive testing.
Moreover, the interpretation of thermal analysis curves requires consideration of the iron’s composition. For hypoeutectic nodular cast iron, a primary austenite plateau may appear, while for hypereutectic compositions, primary graphite precipitation might not yield a distinct plateau due to limited latent heat release. This variability underscores the importance of contextualizing curve features within the specific alloy system. In our study, we focus on near-eutectic nodular cast iron to ensure consistency and clarity in data interpretation.
Experimental Methodology for Thermal Analysis of Nodular Cast Iron
Our experimental approach was designed to isolate the effects of nodularizers and inoculants on the thermal analysis characteristics of nodular cast iron. We conducted two sets of univariate experiments: one varying nodularizers while keeping the inoculant constant, and another varying inoculants while using the optimal nodularizer identified from the first set. This sequential method allows for a systematic evaluation of each additive’s impact.
The base material for all experiments was prepared to mimic the composition of QT400-18 nodular cast iron. The target chemical composition is summarized in the table below:
| Element | Base Iron | Final Iron |
|---|---|---|
| Carbon (C) | 3.7–3.9 | 3.6–3.8 |
| Silicon (Si) | 1.6–1.8 | 2.6–2.8 |
| Manganese (Mn) | <0.3 | <0.3 |
| Magnesium (Mg) | – | 0.035–0.055 |
| Rare Earth (RE) | – | 0.007–0.012 |
| Phosphorus (P) | <0.04 | <0.04 |
| Sulfur (S) | <0.02 | <0.01 |
Raw materials included scrap steel, carbon raiser, ferrosilicon, along with specific nodularizers and inoculants. The nodularizers, labeled Q1 to Q5, contained magnesium and rare earth elements in varying ratios, while the inoculants, labeled Y1 to Y7, were based on ferrosilicon with additions of barium, aluminum, bismuth, strontium, and other elements to enhance nucleation. The detailed compositions of these additives are provided in the following table:
| Additive | Si | Mg | RE Type | Other Elements |
|---|---|---|---|---|
| Q1 | 46 | 7.0 | La65Ce35 | Ca: 2.5, Al: 4.0 |
| Q2 | 46 | 7.0 | La35Ce65 | Ca: 2.5, Al: 1.0 |
| Q3 | 46 | 7.0 | La65Y35 | Ca: 2.5, Bi: 1.0 |
| Q4 | 46 | 7.0 | La35Y65 | Ca: 2.5, Sr: 1.0 |
| Q5 | 46 | 7.0 | La100 | Ca: 2.5, Mn: 3.0 |
| Y1 | 70 | – | La100 | Ba: 1.0, Al: 1.0 |
| Y2 | 76 | – | Ce100 | Ba: 1.0, Al: 1.0 |
| Y3 | 73 | – | Ce100 | Ba: 1.0, Al: 3.0 |
| Y4 | 70 | – | – | Ba: 1.0, Al: 1.0 |
| Y5 | 64 | – | – | Ba: 1.0, Al: 1.0 |
| Y6 | 73 | – | – | Ba: 3.0, Al: 1.0 |
| Y7 | 70 | – | – | Ba: 1.0, Al: 1.0 |
Melting was performed in a laboratory furnace, with each batch weighing 20 kg. The pouring temperature was controlled between 1,510°C and 1,530°C using an infrared thermometer. Treatment involved a sandwich method: nodularizer and inoculant were placed in a preheated ladle pocket, covered with insulating material to minimize oxidation, and then treated with molten iron. After reaction, slag was removed using pearlite. For each experiment, we cast two thermal analysis samples and two shrinkage test samples. Thermal analysis was conducted using thermocouples embedded in standard cups, with data acquisition systems recording temperature-time curves. The characteristic values were extracted automatically via software algorithms.
To assess nodularization, we performed metallographic analysis on the thermal analysis samples. The nodularity rate and graphite count were determined using an image processing algorithm developed in Python. This algorithm segments graphite particles from polished micrographs, calculates their areas and roundness, and classifies them according to ISO 945-4 standards. The nodularity rate \( p_{nod} \) is computed via two methods: count-based and area-based. The formulas are:
Count-based nodularity: $$ p_{nod} = \frac{N_{VI} + N_{V}}{N_{all}} $$ where \( N_{VI} \) and \( N_{V} \) are the counts of type VI and V graphite particles (roundness ≥ 0.6), and \( N_{all} \) is the total count.
Area-based nodularity: $$ p_{nod} = \frac{A_{VI} + A_{V}}{A_{all}} $$ where \( A_{VI} \), \( A_{V} \), and \( A_{all} \) are the corresponding areas.
Shrinkage tendency was evaluated by measuring the volume fraction of porosity in the shrinkage test samples using density measurements or image analysis. The shrinkage rate \( K \) is defined as the percentage of void volume relative to the total sample volume. This comprehensive methodology ensures robust data collection for modeling.
Relationships Between Thermal Analysis Characteristics and Nodularization in Nodular Cast Iron
The analysis of thermal analysis data reveals significant correlations between curve characteristics and the nodularization quality of nodular cast iron. We focus on two key parameters: the eutectic recalescence temperature \(\Delta T\) and the eutectic minimum temperature \(T_{EU}\).
First, the effect of \(\Delta T\) on nodularization is profound. As \(\Delta T\) increases, the nodularity rate and graphite count decrease. This relationship can be expressed as: $$ \text{Nodularity rate} \propto -\Delta T $$ Physically, a larger \(\Delta T\) indicates a faster eutectic reaction, where graphite growth is rapid but less controlled. This leads to irregular graphite shapes and reduced spheroid counts. In contrast, a smaller \(\Delta T\) corresponds to a slower, more orderly eutectic transformation, promoting the formation of spherical graphite with uniform size distribution. Our data, summarized below, confirms this trend across various nodularizers and inoculants:
| Sample Set | ΔT Range (°C) | Average Nodularity Rate (%) | Average Graphite Count (per mm²) |
|---|---|---|---|
| Q1–Q5 (Nodularizers) | 2.85–7.66 | 68.1 | 519.6 |
| Y1–Y7 (Inoculants) | 1.31–11.67 | 70.9 | 458.0 |
For instance, when \(\Delta T = 2.8°C\), the nodularity rate reaches 71.1%, whereas at \(\Delta T = 7.1°C\), it drops to 65.9%. This underscores the importance of controlling recalescence to achieve optimal nodularization in nodular cast iron.
Second, the eutectic minimum temperature \(T_{EU}\) shows a positive correlation with nodularization. Higher \(T_{EU}\) values are associated with increased nodularity rates and graphite counts. This can be modeled as: $$ \text{Nodularity rate} \propto T_{EU} $$ A elevated \(T_{EU}\) suggests reduced undercooling before eutectic solidification, which enhances graphite nucleation and growth. Essentially, a higher \(T_{EU}\) provides more thermal energy for graphite formation, favoring spheroidal over irregular morphologies. Our experimental results demonstrate that as \(T_{EU}\) rises from 1,127.8°C to 1,139.5°C, the nodularity rate improves from 64.7% to 71.1%. This highlights \(T_{EU}\) as a critical indicator for predicting nodularization effectiveness in nodular cast iron.
To quantify these relationships, we performed regression analysis using SPSS. The linear model for nodularity rate \(SG\) based on \(\Delta T\) and \(T_{EU}\) is derived as: $$ SG = 0.532 \times T_{EU} – 0.481 \times \Delta T – 530.937 $$ This equation has a coefficient of determination \(R^2 = 0.85\), indicating a strong fit. The negative coefficient for \(\Delta T\) and positive coefficient for \(T_{EU}\) align with our observational trends. Such a model enables foundries to estimate nodularization quality directly from thermal analysis data, facilitating timely adjustments in processing parameters.
Relationships Between Thermal Analysis Characteristics and Shrinkage Tendency in Nodular Cast Iron
Shrinkage defects in nodular cast iron, including macro-shrinkage cavities and micro-porosity, are closely linked to the time ratios extracted from thermal analysis curves. We examine three ratios: \(G_1/G\) (early shrinkage), \(G_2/G\) (graphitization expansion), and \(G_3/G\) (late shrinkage).
The early shrinkage time ratio \(G_1/G\) represents the period from liquidus to eutectic minimum temperature. During this phase, primary austenite dendrites form, and liquid metal flow compensates for contraction. A longer \(G_1/G\) implies extended dendritic growth, leading to increased feeding resistance and higher shrinkage cavity formation. Our data shows a positive correlation: $$ \text{Shrinkage cavity rate} \propto G_1/G $$ For example, when \(G_1/G = 0.25\), the shrinkage rate measured 3.60%, whereas at \(G_1/G = 0.14\), it was 2.74%. This emphasizes that controlling the early solidification kinetics can mitigate macro-shrinkage in nodular cast iron.
The graphitization expansion time ratio \(G_2/G\) covers the recalescence period where graphite precipitates and expands, counteracting metal contraction. Paradoxically, a longer \(G_2/G\) is associated with higher dispersed shrinkage porosity. This is because prolonged graphitization can isolate liquid pools, hindering feeding and promoting micro-porosity. The relationship is: $$ \text{Dispersed shrinkage rate} \propto G_2/G $$ In our experiments, \(G_2/G\) values ranged from 0.23 to 0.36, with shrinkage porosity rates varying from 2.50% to 3.54%. Thus, optimizing \(G_2/G\) is crucial for minimizing internal porosity in nodular cast iron.
The late shrinkage time ratio \(G_3/G\) corresponds to the final solidification stage. A shorter \(G_3/G\) indicates insufficient time for feeding, exacerbating shrinkage defects. Conversely, a longer \(G_3/G\) allows better compensation. The inverse correlation is: $$ \text{Shrinkage porosity rate} \propto -\frac{1}{G_3/G} $$ Data indicates that when \(G_3/G = 0.48\), shrinkage porosity is 3.60%, but when \(G_3/G = 0.64\), it reduces to 3.06%. Therefore, extending the late solidification phase through process modifications can enhance the integrity of nodular cast iron castings.
To integrate these effects, we developed a nonlinear predictive model for overall shrinkage tendency \(K\) using multiple regression. The model is: $$ K = 28.51 \times G_1 + 30.15 \times G_2^1 – 210.7 \times G_3^1 – 175.63 \times G_2 + 354.23 \times G_2^2 – 13.94 \times G_3 + 20.54 \times G_3^2 + 21.17 $$ where \(G_1\), \(G_2\), and \(G_3\) represent \(G_1/G\), \(G_2/G\), and \(G_3/G\), respectively. This equation has \(R^2 = 0.922\), demonstrating excellent predictive capability. The model accounts for the complex interactions between time ratios and shrinkage, providing a practical tool for defect prevention in nodular cast iron production.
Development and Validation of Predictive Models for Nodular Cast Iron
The core of our research lies in establishing mathematical models that link thermal analysis characteristics to the quality metrics of nodular cast iron. We employed SPSS software for statistical analysis, following a stepwise procedure: correlation analysis, curve estimation, regression modeling, and validation.
For shrinkage tendency prediction, we first computed correlation coefficients among variables. The results are summarized below:
| Variable | \(G_1/G\) | \(G_2/G\) | \(G_3/G\) | Shrinkage Rate |
|---|---|---|---|---|
| \(G_1/G\) | 1.000 | -0.389 | -0.834 | 0.596 |
| \(G_2/G\) | -0.389 | 1.000 | -0.177 | 0.577 |
| \(G_3/G\) | -0.834 | -0.177 | 1.000 | -0.743 |
| Shrinkage Rate | 0.596 | 0.577 | -0.743 | 1.000 |
These correlations guided the selection of predictors for regression. Curve estimation revealed that \(G_1/G\) fits a cubic function, while \(G_2/G\) and \(G_3/G\) fit quadratic functions. The resulting nonlinear model, as presented earlier, was validated against experimental data. The comparison between actual and predicted shrinkage rates shows an average relative error of less than 2.5%, with a maximum error of 4.88%. This confirms the model’s accuracy for practical applications in nodular cast iron foundries.
For nodularization prediction, we focused on \(T_{EU}\) and \(\Delta T\). The correlation analysis yielded coefficients of 0.744 between \(T_{EU}\) and nodularity rate, and -0.693 between \(\Delta T\) and nodularity rate. Linear regression produced the model: $$ SG = 0.532 \times T_{EU} – 0.481 \times \Delta T – 530.937 $$ Validation demonstrated that the predicted nodularity rates deviate from actual values by an average of 2.8%, with all errors below 4%. This model enables real-time assessment of nodularization quality during the casting of nodular cast iron, allowing for immediate corrective actions if needed.
The integration of these models into a quality control system can revolutionize nodular cast iron production. By continuously monitoring thermal analysis curves and computing characteristic values, foundries can predict both shrinkage tendency and nodularization effectiveness before casting solidification completes. This proactive approach reduces scrap, conserves resources, and ensures consistent product quality. Furthermore, our models are adaptable to various grades of nodular cast iron, provided that composition ranges are within the studied limits.
Implications for Industrial Practice and Future Research on Nodular Cast Iron
The findings from this study have direct implications for the foundry industry. Thermal analysis technology, combined with our predictive models, offers a cost-effective solution for quality assurance in nodular cast iron production. Implemented as an inline system, it can provide instant feedback on melt treatment effectiveness, enabling operators to adjust nodularizer or inoculant additions, pouring temperatures, or cooling rates in real time. This not only minimizes defects but also optimizes material usage, contributing to sustainable manufacturing practices.
Specifically, for nodular cast iron, controlling \(\Delta T\) and \(T_{EU}\) through additive selection is crucial. Our research indicates that nodularizers with rare earth combinations like La65Ce35 (Q1) yield lower \(\Delta T\) and higher nodularity, while inoculants containing barium and aluminum (e.g., Y2) enhance graphite counts. By tailoring additives based on thermal analysis data, foundries can achieve desired microstructure outcomes reliably. Additionally, managing time ratios like \(G_1/G\) and \(G_2/G\) through mold design or chilling techniques can mitigate shrinkage defects, improving the yield of sound castings.
Future research should explore the integration of artificial intelligence with thermal analysis for nodular cast iron. Machine learning algorithms could refine our models by incorporating additional variables such as cooling rate, alloy composition, and section thickness. Moreover, extending this methodology to other cast iron grades, like compacted graphite iron, could broaden its applicability. Another avenue is the development of portable thermal analysis devices for small-scale foundries, democratizing access to advanced quality control tools.
In conclusion, our work demonstrates that thermal analysis is a powerful tool for predicting the microstructure and performance of nodular cast iron. The established mathematical models provide accurate forecasts of nodularization and shrinkage tendency, empowering foundries to produce high-quality nodular cast iron components with greater efficiency and consistency. As the demand for durable and reliable cast materials grows, such technological advancements will play a pivotal role in shaping the future of metal casting industries worldwide.