In response to the global emphasis on resource conservation and environmental protection, manufacturing industries are being compelled to enhance energy efficiency and reduce emissions. Among these industries, the sand casting foundry, as a fundamental and widely adopted manufacturing technology, is characterized by high energy consumption and significant pollutant generation. The sand casting foundry process involves injecting molten metal into a sand mold cavity, followed by cooling and finishing to obtain the desired casting. Due to its energy-intensive and polluting nature, the sand casting foundry sector faces an urgent need to improve energy efficiency and minimize environmental impact. Therefore, conducting effective quantitative analysis and optimization of energy consumption and pollutant emissions within sand casting foundry operations is crucial for achieving high energy efficiency and environmental friendliness.
In the manufacturing stage of sand castings, environmental burdens arise from resource consumption and pollutant emissions. It is essential to incorporate potential resource consumption and pollutant emissions into the design basis during the process design phase. In other words, planning should begin at the casting design stage to achieve low-carbon sand casting foundry operations and low greenhouse gas emissions. To calculate and reduce carbon emissions from the process design stage onward, this paper proposes a method for carbon emission calculation and optimization based on process parameters for sand casting foundry. The sand casting foundry process design includes various parameters. According to the different properties of these parameters, they are classified into three categories: independent process parameters, coupled process parameters, and inherent property process parameters. To enable quantitative analysis and optimization of environmental factors such as energy consumption and emissions, the carbon source theory is adopted, and a classification-based calculation method for material carbon sources is proposed. Material carbon sources are further divided into one-time material carbon sources, recyclable material carbon sources, and distributed material carbon sources, leading to the establishment of material carbon source models, energy carbon source models, and undesired carbon source models. Based on the three types of process parameters, calculation models for the three carbon sources are established, and their carbon emissions are computed. The resulting carbon emissions incorporate both direct and indirect emissions from the sand casting foundry process, thereby enabling comprehensive carbon emission modeling.
1. Introduction
The continuous rise of global carbon emissions has attracted worldwide attention. In 2017, global carbon emissions reached 41 billion tons, making climate change a pressing issue. The manufacturing sector contributes over 80% of China’s total carbon emissions, and in 2015, China’s carbon emissions exceeded the combined emissions of the United States and Europe. To promote sustainability, China has set ambitious targets to reduce carbon intensity by 40-50% by 2020 compared to 2005 levels. The “Made in China 2025” strategy emphasizes the importance of green manufacturing and energy efficiency. Casting, as a pillar manufacturing technology, has seen increasing production volumes worldwide. In 2017, China produced more castings than any other country, with output exceeding the total of the next nine largest producing countries combined. Within the casting industry, sand casting foundry remains the dominant method, accounting for 60-70% of all castings produced globally. However, the sand casting foundry is also one of the most energy-intensive processes, consuming 25-30% of the total energy used in the mechanical industry while achieving only 17% energy efficiency. This underscores the critical need for low-carbon research and practice in sand casting foundry operations.
Existing studies have addressed various aspects of sustainable manufacturing and carbon emission reduction. Some researchers have developed models to estimate the carbon footprint of manufacturing processes, while others have optimized process parameters using multi-objective algorithms. In the context of sand casting foundry, several studies have focused on quantifying carbon emissions from different stages or have suggested improvements in energy efficiency. However, most of these methods are not easily applicable during the process design phase, where major decisions about process parameters are made. There is a lack of systematic approaches that link process parameters to carbon emissions in a way that enables designers to perform low-carbon optimization before actual production. This paper aims to fill that gap by proposing a comprehensive, parameter-based carbon emission calculation and optimization framework for sand casting foundry. The proposed method allows designers to estimate emissions early and to adjust process parameters to achieve lower carbon footprints without compromising casting quality.
2. Process Parameter Analysis in Sand Casting Foundry
To build a robust carbon emission model, it is first necessary to understand the process flow and identify the relevant process parameters. A typical sand casting foundry process includes the following operations: sand mixing, molding, core making, mold assembly, melting, pouring, cooling, shakeout, and fettling. A schematic of these operations with inputs (materials, energy) and outputs (products, emissions) is shown in Figure 1 (conceptually). At each stage, energy and materials are consumed, and some are converted into useful products or emitted as waste. The exact amounts of these inputs and outputs are governed by the process parameters selected during the design stage. For example, the pattern design parameters such as machining allowance and draft angle directly influence the amount of metal required, while pouring temperature affects melting energy demand and metal yield. Therefore, process parameters are the key drivers of the environmental performance of a sand casting foundry.
The sand casting foundry process design involves a set of parameters that can be categorized based on their origin and interdependence. I propose to classify them as follows:
- Sand Casting Independent Process Parameters (IP): These are parameters directly determined by the product drawings and the foundry’s production capabilities. Examples include wall thickness, machining allowance, and draft angle. They do not depend on other parameters and can be set independently.
- Sand Casting Coupled Process Parameters (CP): These parameters are determined only after other process parameters are fixed. They depend on the independent parameters and on each other. Examples include pouring temperature, cooling time, and pouring time. For instance, the pouring temperature is influenced by the casting material and the section thickness.
- Sand Casting Inherent Property Process Parameters (IAP): These parameters describe the intrinsic properties of materials used in the process, such as the metal grade and the type of molding sand. They are not chosen arbitrarily but are dictated by the material characteristics and the product requirements.
This classification provides a structured way to map process parameters to carbon sources, as different parameter types influence different carbon flows. For instance, independent parameters mainly affect the quantity of primary materials (e.g., metal) and the energy consumption of handling equipment. Coupled parameters predominantly influence energy consumption during melting and the consumption of auxiliary materials like inoculants. Inherent property parameters determine the type and emission factors of fuels and binders, as well as the gas evolution during pouring.
To illustrate the influence of these parameters, consider the machining allowance. A larger machining allowance results in a heavier casting, which requires more metal to be melted and poured, thereby increasing material consumption and energy use. The draft angle affects the volume of the pattern and the amount of sand required. The fillet radius influences both the pattern geometry and the stress concentration in the casting, which can affect the rejection rate. The pouring temperature directly determines the superheat and thus the energy consumption of the furnace. Therefore, careful selection of these parameters is crucial for minimizing the environmental impact of the sand casting foundry.
3. Carbon Emission Modeling for Sand Casting Foundry
In a sand casting foundry, carbon emissions are generated from three primary sources: material carbon sources, energy carbon sources, and undesired carbon sources. Material carbon sources include all indirect emissions arising from the production and transportation of raw and auxiliary materials, as well as direct emissions from chemical reactions during the process. Energy carbon sources cover both direct emissions from fuel combustion and indirect emissions from the generation of purchased electricity. Undesired carbon sources refer to emissions resulting from the treatment or disposal of waste materials and byproducts, such as sand dust, slag, and defective castings. Each of these sources needs to be modeled separately to enable a comprehensive carbon accounting.
3.1 Material Carbon Source Model
Material carbon sources can be further divided based on the nature of their consumption:
- One-time materials: These are materials that are consumed completely in a single use, such as virgin metal (pig iron, steel scrap), alcohol, and coatings. Their carbon emission is computed as:
$$C_M^{ot} = \sum_{i=1}^{n} m_i \cdot f_i$$
where \( m_i \) is the consumption of the \( i \)-th one-time material, and \( f_i \) is its carbon emission factor.
- Recyclable materials: These include materials like molding sand and return scrap that can be reused multiple times. The emission calculation must account for the material losses during recycling and the energy consumed in the recycling process. The carbon emission from the net consumption is:
$$C_M^{r} = \sum_{j=1}^{m} m_j \cdot (1 – \lambda_j) \cdot f_j$$
where \( m_j \) is the input amount of the \( j \)-th recyclable material, \( \lambda_j \) is its recovery rate, and \( f_j \) is its carbon emission factor.
- Distributed materials: These are materials whose consumption is gradual and long-term, such as steel shot for shot blasting and furnace lining. Their emissions are amortized over the total mass of castings treated during the consumption period. The emission for a single casting is:
$$C_M^{s} = \sum_{k=1}^{h} \left( M_k \cdot \frac{M_{ca}}{M_{aca}} \right) \cdot f_k$$
where \( M_k \) is the total consumption of the \( k \)-th distributed material over a period, \( M_{ca} \) is the mass of the single casting, and \( M_{aca} \) is the total mass of all castings produced during that period.
3.2 Energy Carbon Source Model
Energy consumption in a sand casting foundry includes electricity, natural gas, coal, and other fuels. Electricity usage does not directly release carbon dioxide at the foundry, but the generation of electricity emits CO2 upstream. Therefore, the emission factor for electricity must be used. For fuels, both the upstream production emissions and the direct combustion emissions must be included. The energy carbon emission model is:
$$C_E = E_e \cdot f_E + \sum_{l=1}^{n} \left( E_l \cdot f_{l}^{prod} + E_l \cdot f_{l}^{comb} \right)$$
where \( E_e \) is the total electricity consumption, \( f_E \) is the carbon emission factor for electricity, \( E_l \) is the amount of the \( l \)-th fuel, and \( f_{l}^{prod} \) and \( f_{l}^{comb} \) are the production and combustion emission factors of that fuel, respectively.
3.3 Undesired Carbon Source Model
Undesired carbon sources originate from the treatment of waste products such as sand dust, slag, and waste gas. The treatment processes often consume electricity, which indirectly leads to carbon emissions. The model is:
$$C_U = \sum_{h=1}^{s} M_h \cdot E_h^S \cdot f_E$$
where \( M_h \) is the quantity of the \( h \)-th waste, \( E_h^S \) is the electricity consumption per unit of waste treatment, and \( f_E \) is the electricity emission factor.
3.4 Integration with Process Parameters
The total carbon emission of a sand casting foundry is obtained by summing the contributions from all carbon sources, each expressed as a function of the process parameters. The general formulation is:
$$C = \sum_{i=1}^{n} C_i^{IP} + \sum_{j=1}^{m} C_j^{CP} + \sum_{k=1}^{p} C_k^{IAP}$$
where \( C_i^{IP} \), \( C_j^{CP} \), and \( C_k^{IAP} \) represent the carbon emissions associated with independent, coupled, and inherent property parameters, respectively. Each of these terms can be expanded using the models described above. For example, the independent parameters directly influence the mass of the casting and the mold, thus affecting the material carbon sources. The coupled parameters, particularly pouring temperature, dominate the melting energy consumption. The inherent property parameters determine the type and emission factors of the materials.
To make the model practically usable, we express the relevant quantities as functions of the process parameters. For instance, the mass of the casting can be calculated from the part mass plus the contributions of machining allowance and draft angle, minus the effect of fillet radii. Similarly, the resin sand consumption is a function of the mold dimensions. The melting energy depends on the pouring temperature and the metal mass. By incorporating these relationships into the carbon source models, we obtain a comprehensive process-parameter-based carbon emission model for the sand casting foundry.
4. Low-Carbon Optimization of Sand Casting Foundry Process Parameters
With the carbon emission model established, the next step is to optimize the process parameters to minimize the total carbon emission while ensuring that the casting quality remains acceptable. This optimization problem can be formulated as follows:
The design variables are selected to be the machining allowance \( e \), draft angle \( \alpha \), fillet radius \( r \), and pouring temperature \( T \). These four parameters significantly affect the material and energy consumption, as well as the quality of the casting. The objective function is the total carbon emission \( C \) expressed as \( F(e, \alpha, r, T) \).
The constraints are derived from the practical ranges of these parameters as recommended by sand casting design standards and foundry practice. For example, the machining allowance must be within a minimum and maximum bound depending on the casting size and material. The draft angle should not exceed a certain limit to avoid affecting the dimensional accuracy. The fillet radius must be larger than a minimum value to prevent stress concentration. The pouring temperature should be between the liquidus temperature plus minimum superheat and the liquidus temperature plus maximum superheat.
Thus, the optimization model is:
$$\min \quad F(e, \alpha, r, T) = C$$
$$\text{s.t.} \quad e_{\min} \le e \le e_{\max}, \quad \alpha_{\min} \le \alpha \le \alpha_{\max}, \quad r_{\min} \le r \le r_{\max}, \quad T_l + \Delta T_{\min} \le T \le T_l + \Delta T_{\max}$$
To solve this nonlinear optimization problem, a genetic algorithm (GA) is adopted due to its robustness and ability to handle complex, multi-variable problems. The genetic algorithm operates on a population of candidate solutions, where each individual encodes the four variables. The fitness of each individual is evaluated using the objective function (or its transformed form). The algorithm iteratively applies selection, crossover, and mutation operations to evolve the population toward an optimal solution. The typical steps are:
- Encoding: The variables are encoded as binary strings. For instance, the machining allowance \( e \) in the range [3.5, 9] mm can be represented by a fixed number of bits, e.g., 4 bits giving 16 levels. The pouring temperature \( T \) in the range [1440, 1450] °C is encoded similarly.
- Initialization: A random initial population of size \( N \) (e.g., 100) is generated.
- Fitness evaluation: The carbon emission is calculated for each individual using the comprehensive model. For minimization, the fitness is defined as \( F_{fit} = C_{max} – C \) if \( C < C_{max} \), otherwise zero.
- Selection: A stochastic tournament selection is used to choose parents for reproduction.
- Crossover: Two-point crossover is applied with a probability \( p_c \) (e.g., 0.8) to produce offspring.
- Mutation: Each bit in the offspring chromosomes is flipped with a small probability \( p_m \) (e.g., 0.01) to maintain genetic diversity.
- Termination: The algorithm is terminated after a predefined number of generations (e.g., 150) or when the fitness no longer improves significantly.
The outcome of the genetic algorithm is a set of optimized process parameters that yield the lowest carbon emission while satisfying all constraints. The obtained solution serves as the basis for the improved process plan.
5. Case Study and Application
To demonstrate the effectiveness of the proposed methodology, a case study is conducted for an electric motor housing produced in a sand casting foundry. The foundry receives a product drawing for a motor housing with a mass of 579.69 kg, to be cast in HT250 gray iron. The production line uses a self-hardening resin sand system, and the sand recovery rate is 96.25%. The main equipment and their power ratings are listed in Table 1. The process steps and their duration for producing the housing are shown in Table 2 along with the equipment power. Also, the electric energy required for melting one ton of iron at different pouring temperatures is presented in Table 3.
| Equipment | Power (kW) |
|---|---|
| Sand mixer | 11.5 |
| Roller conveyor | 2.2 |
| Drying furnace | 104.2 |
| Compaction table | 3.0 |
| Coating machine | 1.85 |
| Bridge pattern drawing machine | 6.6 |
| Hardener proportioning device | 1.5 |
| Sand reclamation line | 178.5 |
| Process step | Duration (s) | Equipment power (kW) |
|---|---|---|
| Sand mixing | 1800 | 11.5 |
| Molding and filling | 1200 | 3+2.2 |
| Pattern drawing | 600 | 6.6 |
| Mold drying | 3600 | 104.2 |
| Coating | 1200 | 1.85 |
| Mold assembly | 600 | 6.6 |
| Pouring | 20 | — |
| Cooling | 86400 | — |
| Shakeout | 600 | 6.6 |
| Fettling | 1800 | — |
| Shot blasting | 1800 | — |
| Pouring temperature (°C) | Energy per ton (kW·h) |
|---|---|
| 1415 | 485 |
| 1420 | 510 |
| 1425 | 535 |
| 1430 | 560 |
| 1435 | 585 |
| 1440 | 610 |
| 1445 | 635 |
| 1450 | 660 |
The carbon emission factors for various materials and energy carriers are listed in Table 4. These factors are based on national databases and published literature. Note that the emission factor for electricity includes the upstream emissions from power generation. For fuels, the factor accounts for both production and combustion unless otherwise stated.
| Material/Fuel | Emission factor (kgCO2/kg or kgCO2/kWh) |
|---|---|
| Electricity | 0.93 |
| Coating | 6.023 |
| Natural gas | 2.162 |
| Alcohol | 0.48 |
| Silica sand | 0.0254 |
| Pig iron | 2.13 |
| Steel scrap | 8.2 |
| Tap water | 0.194 |
5.1 Initial Process Plan
The initial process plan (Plan 1) is designed using conventional rules and experience. The key process parameters are listed in Table 5. The mold dimensions for the motor housing are 1350 mm × 1250 mm × 550 mm for both cope and drag. The casting is produced with one cavity per mold. The pouring temperature is set at 1446°C. The gas evolution of the resin sand is 15.5 ml/g. The cooling time is 24 hours.
| Parameter | Plan 1 | Plan 2 |
|---|---|---|
| Draft angle | 2° | 0.42° |
| Machining allowance | 9 mm | 3.5 mm |
| Fillet radius | 8 mm | 10 mm |
| Pouring temperature | 1446 °C | 1440 °C |
| Pouring cup cross-section | 70 mm | 70 mm |
| Runner cross-section | 40/50×70 mm | 40/50×70 mm |
| Vent cross-section | Φ110×300 mm | Φ110×300 mm |
| Gas evolution of sand | 15.5 ml/g | 15.5 ml/g |
| Cooling time | 24 h | 24 h |
Based on these parameters, the mass of the casting is calculated. The volume of machining allowance is computed from the allowance thickness and the surface areas. The draft angle contributes an additional volume depending on the vertical heights and the tangent of the angle. The fillet radius reduces the sharp corners but adds a small volume. The net casting mass becomes:
$$M_C = M_P + (V_e + V_{\alpha} – V_r) \cdot \rho_{HT250}$$
For Plan 1, the calculated volumes are:
$$V_e = 9\,\text{mm} \times [\pi(300.99^2 – 247.65^2) + \pi(410.21^2 – 360.68^2)] = 1907008.80 \,\text{mm}^3$$
$$V_{\alpha} = \pi \times 0.5 \times |\tan 2^\circ| \times [(300.99^2 – 247.65^2)\times 300.99 + (410.21^2 – 360.68^2)\times 410.21] = 1342319.15 \,\text{mm}^3$$
$$V_r = (\pi \times 8^2 – 8^2) \times 1752.98 = 240266.82 \,\text{mm}^3$$
Therefore, the casting mass is \( 579.69 + (1907008.80 + 1342319.15 – 240266.82) \times 7.2 \times 10^{-6} = 601.36 \) kg. The gating system mass is 135 kg, and the total metal weight is 736.36 kg. The amount of resin sand required is 2747.09 kg.
Using the carbon emission models, the emissions for Plan 1 are computed as follows:
Material carbon source (IP):
$$C_M^{IP} = (75.33 \times 2.13 + 378.04 \times 8.2 \times 0.5 + 8.54 \times 0.855) + [2747.09 \times (1-0.9625) \times 0.2543] + (1.35 \times 3.6318) = 1748.83 \,\text{kgCO}_2$$
Energy carbon source (IP):
$$C_E^{IP} = (98.85 + 53.94) \times 0.93 = 142.10 \,\text{kgCO}_2$$
Undesired carbon source (IP):
$$C_U^{IP} = 0.21 \times (36.82 + 147.27 + 0.002) \times 10^{-3} = 0.038 \,\text{kgCO}_2$$
Material carbon source (CP):
$$C_M^{CP} = 0.29 \times 0.194 + 0.08 \times 0.0428 + 736.36 \times 0.008 \times 8.863 = 52.27 \,\text{kgCO}_2$$
Energy carbon source (CP):
$$C_E^{CP} = 736.36 \times 0.62 \times 0.93 = 422.53 \,\text{kgCO}_2$$
Note: 0.62 kW·h/kg is the melting energy per kg of metal at 1446°C, interpolating the trend from Table 3.
Undesired carbon source (CP):
$$C_U^{CP} = 736.36 \times 3.69 \times 10^{-3} = 2.72 \,\text{kgCO}_2$$
Material carbon source (IAP):
$$C_M^{IAP} = 2.28 \times 6.02 + 2 \times 4 + 2 \times 6.66 = 35.10 \,\text{kgCO}_2$$
Energy carbon source (IAP):
$$C_E^{IAP} = 1.49 \times (0.48 + 1.913) + 11.12 = 14.68 \,\text{kgCO}_2$$
Undesired carbon source (IAP):
$$C_U^{IAP} = 15.5 \times 2747.10 \times 0.0431 \times 1.96 \times 10^{-3} = 3.60 \,\text{kgCO}_2$$
The total carbon emission for Plan 1 is obtained by summing all terms:
$$C_{total}^{Plan1} = (1748.83+142.10+0.04) + (52.27+422.53+2.72) + (35.10+14.68+3.60) = 2421.86 \,\text{kgCO}_2$$
The breakdown by process parameter type and carbon source is given in Table 6 and Table 7.
| Carbon source | IP | CP | IAP |
|---|---|---|---|
| Material | 1748.83 | 52.27 | 35.10 |
| Energy | 142.10 | 422.53 | 14.68 |
| Undesired | 0.04 | 2.72 | 3.60 |
| Total | 1890.97 | 477.52 | 53.38 |
| Carbon source | Material | Energy | Undesired |
|---|---|---|---|
| IP | 1748.83 | 142.10 | 0.04 |
| CP | 52.27 | 422.53 | 2.72 |
| IAP | 35.10 | 14.68 | 3.60 |
| Total | 1836.20 | 579.31 | 6.36 |
5.2 Optimization Using Genetic Algorithm
The genetic algorithm is implemented in MATLAB R2016b. The initial population size is set to 100, the maximum number of generations is 150, the elite probability is 0.05, the crossover probability is 0.8, and the mutation probability is 0.01. The optimization variables are constrained as follows:
- Machining allowance: \( 3.5 \le e \le 9 \) mm (from JB/T 5105-1991)
- Draft angle: \( 0.42 \le \alpha \le 3 \) degrees
- Fillet radius: \( 4 \le r \le 10 \) mm
- Pouring temperature: \( 1440 \le T \le 1450 \) °C (based on foundry practice)
The objective function is the total carbon emission \( C \) as defined by the comprehensive model. The genetic algorithm iterates to minimize this value. Figure 1 presents the fitness convergence history, and Figure 2 shows the best individual values over generations. The optimal solution found after 150 generations is:
$$[e, \alpha, r, T] = [3.5, 0.42, 10, 1440]$$
These values correspond to a machining allowance of 3.5 mm, a draft angle of 0.42 degrees, a fillet radius of 10 mm, and a pouring temperature of 1440 °C. This optimized parameter set is used to construct Plan 2.
5.3 Optimized Process Plan (Plan 2)
Using the optimized parameters, the casting mass is recalculated:
$$V_e = 3.5 \times [\pi(300.99^2 – 247.65^2) + \pi(410.21^2 – 360.68^2)] = 741614.53 \,\text{mm}^3$$
$$V_{\alpha} = \pi \times 0.5 \times |\tan 0.42^\circ| \times [(300.99^2 – 247.65^2)\times 300.99 + (410.21^2 – 360.68^2)\times 410.21] = 281777.57 \,\text{mm}^3$$
$$V_r = (\pi \times 10^2 – 10^2) \times 1752.98 = 375416.91 \,\text{mm}^3$$
The casting mass becomes:
$$M_C = 579.69 + (741614.53 + 281777.57 – 375416.91) \times 7.2 \times 10^{-6} = 584.36 \,\text{kg}$$
The gating system remains at 135 kg, so the total metal weight is 719.36 kg. The resin sand requirement is 2747.10 kg (unchanged because the mold size is the same).
Now, the carbon emissions for Plan 2 are calculated using the same models:
Material carbon source (IP):
$$C_M^{IP} = (73.59 \times 2.13 + 369.32 \times 8.2 \times 0.5 + 8.34 \times 0.855) + [2747.10 \times 0.0375 \times 0.2543] + (1.31 \times 3.6318) = 1709.04 \,\text{kgCO}_2$$
Energy carbon source (IP):
$$C_E^{IP} = (98.85 + 53.94) \times 0.93 = 142.10 \,\text{kgCO}_2$$
The IP-based energy consumption is unaffected by the changed parameters because the handling equipment is the same. However, the melting energy is reduced due to the lower pouring temperature (1440°C). The melting energy per kg at 1440°C is 0.596 kW·h/kg (from Table 3 interpretation). Therefore:
Energy carbon source (CP):
$$C_E^{CP} = 719.36 \times 0.596 \times 0.93 = 398.74 \,\text{kgCO}_2$$
Material carbon source (CP):
$$C_M^{CP} = 0.28 \times 0.194 + 0.08 \times 0.0428 + 719.36 \times 0.008 \times 8.863 = 51.06 \,\text{kgCO}_2$$
Undesired carbon source (CP):
$$C_U^{CP} = 719.36 \times 3.69 \times 10^{-3} = 2.65 \,\text{kgCO}_2$$
Material carbon source (IAP):
$$C_M^{IAP} = 2.21 \times 6.02 + 2 \times 4 + 2 \times 6.66 = 34.70 \,\text{kgCO}_2$$
Energy carbon source (IAP):
$$C_E^{IAP} = 1.46 \times (0.48 + 1.913) + 10.86 = 14.32 \,\text{kgCO}_2$$
Undesired carbon source (IAP):
$$C_U^{IAP} = 15.5 \times 2747.10 \times 0.0431 \times 1.96 \times 10^{-3} = 3.60 \,\text{kgCO}_2$$
The total carbon emission for Plan 2 is:
$$C_{total}^{Plan2} = (1709.04+142.10+0.04) + (51.06+398.74+2.65) + (34.70+14.32+3.60) = 2356.24 \,\text{kgCO}_2$$
The detailed breakdown for Plan 2 is given in Table 8 and Table 9.
| Carbon source | IP | CP | IAP |
|---|---|---|---|
| Material | 1709.04 | 51.06 | 34.70 |
| Energy | 142.10 | 398.74 | 14.32 |
| Undesired | 0.04 | 2.65 | 3.60 |
| Total | 1851.18 | 452.45 | 52.62 |
| Carbon source | Material | Energy | Undesired |
|---|---|---|---|
| IP | 1709.04 | 142.10 | 0.04 |
| CP | 51.06 | 398.74 | 2.65 |
| IAP | 34.70 | 14.32 | 3.60 |
| Total | 1794.80 | 555.16 | 6.29 |
5.4 Comparison and Discussion
Comparing the two plans, the optimized Plan 2 achieves a reduction in total carbon emission of:
$$\Delta C = 2421.86 – 2356.24 = 65.62 \,\text{kgCO}_2 \quad (2.71\%)$$
The relative reductions for each carbon source and parameter type are summarized in Table 10 and Table 11.
| Carbon source | Plan 1 | Plan 2 | Reduction | Percentage |
|---|---|---|---|---|
| Material | 1836.20 | 1794.80 | 41.40 | 2.25% |
| Energy | 579.31 | 555.16 | 24.15 | 4.17% |
| Undesired | 6.36 | 6.29 | 0.07 | 1.10% |
| Parameter type | Plan 1 | Plan 2 | Reduction | Percentage |
|---|---|---|---|---|
| IP | 1890.97 | 1851.18 | 39.79 | 2.10% |
| CP | 477.52 | 452.45 | 25.07 | 5.25% |
| IAP | 53.38 | 52.62 | 0.76 | 1.42% |
The results indicate that the energy carbon source shows the largest percentage reduction (4.17%), primarily resulting from the decrease in pouring temperature, which lowers the melting energy demand. The material carbon source exhibits a smaller percentage reduction (2.25%), attributable to the reduction in machining allowance and draft angle, which reduce the metal mass. The undesired carbon source is only marginally reduced, but its environmental toxicity makes it an important area for future improvement. From the parameter-type perspective, the coupled parameters (CP) show the largest relative reduction (5.25%) because they directly control melting energy. The independent parameters (IP) contribute the largest absolute reduction (39.79 kgCO2). The inherent property parameters (IAP) show minimal change since they are mostly fixed by material selection.
In terms of the proportion of carbon sources, material carbon remains the dominant contributor (75.8% in Plan 1 and 76.2% in Plan 2). Energy carbon accounts for about 23-24% of the total, while undesired carbon is negligible in quantity but has significant environmental hazard potential. The melting energy constitutes about 72-73% of the energy carbon emission, which matches previous industry statistics. Therefore, optimizing the melting process is a key lever for further emission reduction. Additionally, reducing material waste by optimizing process parameters offers a promising path to lower the largest share of emissions.
The quality of the optimized casting was verified using ProCAST simulation software. The simulation indicated that the casting completely solidifies in the mold, and only minor shrinkage is present in the gating system and at the base of the housing, which is within acceptable limits. Thus, the optimized parameters do not compromise the casting quality. This confirms that the genetic algorithm-based approach can effectively reduce carbon emissions while maintaining product integrity.
The proposed methodology offers several advantages. First, it enables the estimation of carbon emissions during the process design phase, providing a forward-looking assessment. Second, it allows the optimization of process parameters before production, avoiding expensive retrofits. Third, it supports design engineers in creating low-carbon casting processes. The method is particularly useful for sand casting foundry enterprises that wish to benchmark their environmental performance and identify reduction opportunities. The case study demonstrates that a mere adjustment of four key parameters can lead to a 2.71% reduction in total carbon emission. For a large foundry producing 40,000 tons of castings per year, this corresponds to an annual reduction of 1,084 tons of CO2. Thus, the method has significant industrial applicability.
6. Conclusion and Future Work
This paper has presented a comprehensive methodology for carbon emission calculation and optimization in a sand casting foundry, based on process parameters. The main contributions are as follows:
- An analysis of sand casting foundry process parameters led to their classification into three types: independent, coupled, and inherent property parameters. This classification facilitates the mapping of parameters to carbon sources.
- Carbon source models for materials, energy, and undesired byproducts were established, with material carbon sources further segmented into one-time, recyclable, and distributed categories. These models were integrated with the process parameters to create a full carbon emission calculation framework for the sand casting foundry.
- A low-carbon optimization model using a genetic algorithm was developed to determine the optimal machining allowance, draft angle, fillet radius, and pouring temperature that minimize total carbon emissions while satisfying quality constraints.
- A case study on a motor housing was conducted to validate the method. The optimization resulted in a 2.71% reduction in total carbon emissions, with the energy carbon source experiencing the largest relative decrease (4.17%). The coupled parameters showed a 5.25% reduction, demonstrating the importance of melting temperature optimization.
These findings confirm that the proposed method is an effective tool for reducing environmental impacts in sand casting foundry operations from the design stage. It enables designers to proactively choose process parameters that lead to lower carbon footprints. Future research should expand the scope to include the design of gating systems and risers, as these also affect material and energy efficiency. Moreover, the relationship between process parameters and casting defects could be investigated to predict and prevent quality issues during the design phase. The current study focuses on a single cavity mold; extending the model to multi-cavity molds and different sand systems would further enhance its applicability. Nonetheless, the presented methodology provides a solid foundation for achieving greener sand casting foundry practices.