In this dissertation, I present a comprehensive study on the calculation and optimization of carbon emissions in sand casting processes, focusing on the role of process parameters. The research is motivated by the urgent global demand for sustainable manufacturing, particularly in energy-intensive and high-emission industries such as sand casting. I analyze the sand casting process from a carbon emission perspective, classify process parameters into independent, coupled, and inherent property parameters, and establish corresponding carbon emission models for material, energy, and undesired carbon sources. A low-carbon optimization method based on genetic algorithms is developed to optimize key process parameters while ensuring casting quality. A real-world case study of a motor housing produced by sand casting is conducted to validate the proposed models and optimization approach. The results demonstrate that the optimized process parameters can reduce total carbon emissions by 2.71% without introducing any sand casting defects. The proposed method enables foundry engineers to estimate and reduce carbon emissions at the process design stage, thereby contributing to the goal of green foundry and sustainable manufacturing.
1. Introduction
Over the past century, global climate change has become a prominent issue, and carbon emissions have been identified as the primary driver of global warming. Manufacturing activities are among the largest contributors to carbon emissions, accounting for nearly one-third of global energy consumption and 36% of carbon dioxide emissions. In China, the manufacturing sector contributes to more than 80% of the national carbon emissions. In 2017, global carbon emissions reached 41 billion tons, prompting governments and industries to seek effective measures for emission reduction. The Chinese government has set ambitious targets, such as reducing carbon intensity per unit of GDP by 40-50% by 2020 compared to 2005 levels. The “Made in China 2025” strategy emphasizes green manufacturing and energy efficiency as key priorities.
Among various manufacturing processes, casting is a foundational technology for producing metal components. Sand casting, in particular, is the most widely used casting method, accounting for 60-70% of the world’s total casting output. However, sand casting is also one of the most energy-intensive and environmentally harmful manufacturing processes. In China, the casting industry consumes 25-30% of the total energy used in the machinery industry, while its energy utilization rate is only about 17%. The production of one ton of sand castings generates approximately 50 kg of dust, 200 kg of waste slag, 20 m³ of exhaust gas, and 1.3 to 1.5 tons of waste sand. These figures highlight the urgent need to improve energy efficiency and reduce pollutant emissions in sand casting.
Existing research on carbon emissions in sand casting has primarily focused on calculating emissions from different process stages or proposing energy-saving measures. However, most of these methods are applied after the process design has been finalized, leaving limited room for optimization. To achieve significant emission reductions, it is essential to incorporate carbon emission considerations into the early stage of process design. This requires a systematic understanding of how sand casting process parameters influence carbon emissions. In this thesis, I propose a process-parameter-based carbon emission calculation and optimization method for sand casting. By classifying process parameters and establishing quantitative relationships between these parameters and carbon sources, I enable foundry designers to predict and reduce emissions before actual production.
The main contributions of this work are as follows. First, I analyze the sand casting process and classify process parameters into three categories: independent process parameters (IP), coupled process parameters (CP), and inherent attribute process parameters (IAP). Second, I develop carbon emission models for material, energy, and undesired carbon sources, and integrate them into a unified carbon emission calculation framework based on the proposed parameter classification. Third, I formulate a low-carbon optimization model with process parameters as decision variables and use a genetic algorithm to solve it. Finally, I demonstrate the effectiveness of the proposed method through a case study of a motor housing, confirming that optimized process parameters can reduce carbon emissions while maintaining casting quality and avoiding sand casting defects.
2. Analysis of Sand Casting Process Parameters
2.1 Sand Casting Process Description
A typical sand casting process consists of several steps: sand mixing, molding, core making, assembling, melting, pouring, cooling, shakeout, and finishing. Each step consumes energy and materials, and generates emissions and waste. The relationship between the process steps and carbon sources is illustrated through the material flow, energy flow, and waste flow. For instance, the sand mixing step uses a sand mixer to blend silica sand, resin, and hardener. The molding step fills the sand into the flask, compacts it, and sometimes cures it in an oven. The melting step consumes a significant amount of electricity to melt the metal. The pouring step introduces molten metal into the mold cavity. The cooling step allows the casting to solidify. The shakeout step removes the sand mold from the casting. The finishing step includes shot blasting and grinding to remove burrs and surface imperfections. Each of these steps can be linked to specific carbon sources, which are affected by the choice of process parameters.
2.2 Classification of Process Parameters
Sand casting process design involves both process scheme design and process parameter design. The process scheme design includes selecting molding methods, determining parting surfaces, designing gating systems, feeding systems, and chill placement. Process parameters include casting shrinkage, machining allowance, draft angle, minimum cast hole size, process correction amount, parting negative allowance, core negative allowance, anti-deformation amount, pouring temperature, pouring time, and material grade. In this thesis, I focus on the process parameters that can be quantitatively modeled. Based on their characteristics, I classify sand casting process parameters into three types:
| Parameter Type | Definition | Examples |
|---|---|---|
| Independent Process Parameters (IP) | Parameters directly determined by the product drawing and the foundry’s production conditions. | Machining allowance, draft angle, wall thickness, casting dimensions. |
| Coupled Process Parameters (CP) | Parameters that are determined based on other already-defined parameters and process schemes. | Pouring temperature, cooling time, melting temperature. |
| Inherent Attribute Process Parameters (IAP) | Parameters that represent the intrinsic properties of the materials involved in the casting process. | Metal grade, sand type, binder type. |
This classification is essential for establishing the mapping between process parameters and carbon emissions. The independent parameters directly influence the amount of primary materials such as iron and steel. The coupled parameters mainly affect energy consumption during melting and the amount of auxiliary materials. The inherent attribute parameters determine the carbon emission factors of the materials and the gas evolution characteristics of the sand mold.
3. Carbon Emission Modeling for Sand Casting
3.1 Classification of Carbon Sources
In the sand casting process, carbon emissions arise from three main sources: material carbon sources, energy carbon sources, and undesired carbon sources. Material carbon sources include emissions from the production and transportation of raw materials (e.g., pig iron, scrap steel, sand) as well as direct emissions from chemical reactions during material use. Energy carbon sources include emissions from the consumption of electricity and fuels, including both direct emissions from fuel combustion and indirect emissions from the production of energy carriers. Undesired carbon sources refer to emissions from the processing and disposal of waste materials such as sand dust, slag, and waste gas, as well as direct emissions from thermal decomposition of binders.
Material carbon sources are further divided into three subcategories based on their usage characteristics:
- One-time material carbon sources: Materials that are consumed directly and irreversibly in a single process, such as pig iron, scrap steel, alcohol, and coatings.
- Recyclable material carbon sources: Materials that can be recovered and reused, such as return sand and return iron. The carbon emission calculation must account for the recovery rate and the energy consumed during recovery.
- Shared material carbon sources: Materials that are consumed slowly over a long period and are apportioned to each casting, such as steel shot and furnace linings.
3.2 Carbon Emission Models for Different Carbon Sources
The carbon emission from one-time materials is calculated as follows:
$$C_{M}^{ot} = \sum_{i=1}^{n} m_i \cdot f_i$$
where mi is the consumption of the i-th one-time material and fi is its carbon emission factor.
The carbon emission from recyclable materials is calculated as:
$$C_{M}^{r} = \sum_{j=1}^{m} m_j \cdot (1 – \lambda_j) \cdot f_j$$
where mj is the input amount of the j-th recyclable material, λj is its recovery rate, and fj is its carbon emission factor.
The carbon emission from shared materials apportioned to a single casting is:
$$m_k = M_k \cdot \frac{M_{ca}}{M_{aca}}$$
$$C_{M}^{s} = \sum_{k=1}^{h} m_k \cdot f_k = \sum_{k=1}^{h} M_k \cdot \frac{M_{ca}}{M_{aca}} \cdot f_k$$
where Mk is the total consumption of the k-th shared material in a certain period, Mca is the mass of the single casting, and Maca is the total mass of all castings produced during that period.
The energy carbon emission model considers both electricity and fuels:
$$C_E = E_e \cdot f_E + \sum_{l=1}^{n} (E_l \cdot f_l + E_l \cdot f_f)$$
where Ee is the total electricity consumption, fE is the carbon emission coefficient of electricity, El is the amount of the l-th fuel, fl is the carbon emission factor for fuel production, and ff is the carbon emission factor for fuel combustion.
The undesired carbon emission model is:
$$C_U = \sum_{h=1}^{n} M_h \cdot E_h^S \cdot f_E$$
where Mh is the amount of the h-th waste material, EhS is the electricity consumption per unit of waste treatment, and fE is the electricity carbon emission coefficient.
3.3 Carbon Emission Calculation Based on Process Parameters
Based on the classification of process parameters (IP, CP, IAP), I establish the following calculation formulas for each carbon source.
3.3.1 Carbon Emissions Based on Independent Process Parameters (IP)
The material carbon emission based on IP is:
$$C_{M}^{IP} = \sum_{a=1}^{n} F_a^{ot}(x_{IP}) \cdot f_a + \sum_{b=1}^{n} F_b^{r}(x_{IP}) \cdot [(1-\lambda_b) f_b] + \sum_{c=1}^{n} F_c^{s}(x_{IP}) \cdot f_c$$
where Faot(xIP) is the amount of the a-th one-time material as a function of independent parameters, Fbr(xIP) is the amount of the b-th recyclable material, and Fcs(xIP) is the apportioned amount of the c-th shared material.
The energy carbon emission based on IP is:
$$C_{E}^{IP} = \sum_{d=1}^{n} f_d^{E}(x_{IP}) \cdot f_E + \sum_{e=1}^{n} F_e^{r}(x_{IP}) \cdot E_e^{r} \cdot f_E$$
where fdE(xIP) is the power consumption of equipment d, and Eer is the electricity consumption for recycling the e-th material.
The undesired carbon emission based on IP is:
$$C_{U}^{IP} = \sum_{g=1}^{n} \lambda_g^{ot}(x_{IP}) \cdot E_g \cdot f_E + \sum_{o=1}^{n} \lambda_o^{r}(x_{IP}) \cdot E_o \cdot f_E + \sum_{f=1}^{n} \lambda_f^{s}(x_{IP}) \cdot E_f \cdot f_E$$
where λgot, λor, and λfs are the generation coefficients of undesired sources for one-time, recyclable, and shared materials, respectively, and Eg, Eo, Ef are the corresponding treatment electricity consumptions.
3.3.2 Carbon Emissions Based on Coupled Process Parameters (CP)
The material carbon emission based on CP is:
$$C_{M}^{CP} = \sum_{q=1}^{n} F_q^{ot}(x_{CP}) \cdot f_q$$
where Fqot(xCP) represents the consumption of auxiliary materials such as coolant water, thermocouples, release agents, and inoculants.
The energy carbon emission based on CP is:
$$C_{E}^{CP} = E_{se}(x_{CP}) \cdot F_{se}^{r}(x_{IP}) \cdot f_E$$
where Ese(xCP) is the electricity consumption per ton of molten metal, and Fser(xIP) is the metal weight.
The undesired carbon emission based on CP is:
$$C_{U}^{CP} = \sum_{s=1}^{n} F_s(x_{CP}) \cdot a_s \cdot F_{se}^{r}(x_{IP})$$
where Fs(xCP) is the unit carbon emission of the s-th undesired source (e.g., slag from melting), and as is its proportion.
3.3.3 Carbon Emissions Based on Inherent Attribute Process Parameters (IAP)
The material carbon emission based on IAP is:
$$C_{M}^{IAP} = \sum_{t=1}^{n} F_t^{ot}(x_{IAP}) \cdot f_t$$
where Ftot(xIAP) denotes the consumption of materials such as coatings and filters, which depend on the sand type and metal grade.
The energy carbon emission based on IAP is:
$$C_{E}^{IAP} = \sum_{l=1}^{n} F_l^{El}(x_{IAP}) \cdot (f_l + f_f)$$
where FlEl(xIAP) is the amount of fuel (e.g., alcohol, natural gas) used, which depends on the inherent properties of the sand and metal.
The undesired carbon emission based on IAP is:
$$C_{U}^{IAP} = F(x_{IAP}) \cdot V_s \cdot f_g$$
where F(xIAP) is the gas evolution per unit volume of sand, Vs is the sand volume, and fg is the carbon emission factor of the gas.
3.4 Total Carbon Emission Model
Combining all the above sub-models, the total carbon emission of the sand casting process based on process parameters is:
$$C = C_{M}^{IP} + C_{E}^{IP} + C_{U}^{IP} + C_{M}^{CP} + C_{E}^{CP} + C_{U}^{CP} + C_{M}^{IAP} + C_{E}^{IAP} + C_{U}^{IAP}$$
This comprehensive model incorporates both direct and indirect carbon emissions from the entire sand casting process. It enables designers to quantitatively evaluate the carbon emission impact of different process parameter choices.
4. Low-Carbon Optimization of Sand Casting Process Parameters
4.1 Optimization Variables
Based on the carbon emission model, I select four key process parameters as optimization variables: machining allowance e, draft angle α, casting fillet radius r, and pouring temperature T. These variables significantly affect material consumption, energy consumption, and hence carbon emissions. They are also constrained by casting quality requirements to avoid sand casting defects such as shrinkage, misrun, and hot tears.
4.2 Optimization Objective Function
The objective is to minimize the total carbon emission function:
$$\min F[e, \alpha, r, T] = \min C$$
given the process conditions and constraints.
4.3 Constraints
- Machining allowance constraint: The machining allowance must lie within a feasible range determined by casting dimensions, material, and molding conditions:
$$\{cond_1,\ldots,cond_n | a_{min}\} \le a \le \{cond_1,\ldots,cond_n | a_{max}\}$$
- Draft angle constraint: The draft angle must not exceed a maximum value to ensure easy pattern withdrawal without damaging the mold:
$$\alpha \le \{cond_1,\ldots,cond_m | \alpha_{max}\}$$
- Fillet radius constraint: The fillet radius must be within limits to prevent stress concentration and sand casting defects:
$$\{cond_1,\ldots,cond_l | r_{min}\} < r < \{cond_1,\ldots,cond_l | r_{max}\}$$
- Pouring temperature constraint: The pouring temperature must be high enough to ensure complete filling but low enough to avoid defects:
$$T_l + \Delta T_{min} \le T_c \le T_l + \Delta T_{max}$$
where Tl is the liquidus temperature of the alloy, and ΔTmin and ΔTmax are the minimum and maximum superheat values.
4.4 Genetic Algorithm-Based Solution
I use a genetic algorithm (GA) to solve the optimization problem. The GA mimics natural selection and evolution. The key steps are:
- Encoding: Each set of process parameters is encoded as a binary chromosome, with each parameter represented by a gene segment.
- Initial population: A random initial population of N individuals is generated.
- Fitness evaluation: The fitness of each individual is evaluated using the objective function. Since I minimize the total carbon emission, the fitness is transformed as:
$$Fit(x) = \begin{cases} C_{max} – f(x) & \text{if } f(x) < C_{max} \\ 0 & \text{otherwise} \end{cases}$$
- Selection: I use stochastic tournament selection to choose individuals for reproduction.
- Crossover: Two-point crossover is applied to create offspring.
- Mutation: Bit-flip mutation is applied with a small probability to maintain diversity.
- Termination: The algorithm terminates after a predefined number of generations or when the fitness converges.
The GA parameters used in this study are given in the following table:
| Parameter | Value |
|---|---|
| Population size | 100 |
| Maximum generations | 150 |
| Elite count | 5 |
| Crossover probability | 0.8 |
| Mutation probability | 0.01 |
5. Case Study and Application
5.1 Sand Casting Process Conditions
I applied the proposed method to a real sand casting foundry that produces motor housings. The motor housing material is HT250 gray cast iron, with a net weight of 579.69 kg. The foundry uses self-hardening resin sand for molding. The sand reclamation rate is 96.25%. The rated powers of the main equipment are: sand mixer 11.5 kW, roller conveyor 2.2 kW, drying oven 104.2 kW, compaction table 3 kW, coating machine 1.85 kW, bridge-type pattern drawing machine 6.6 kW, hardener proportioning instrument 1.5 kW, and sand handling line 178.5 kW. The flask internal dimensions are 1350 mm × 1250 mm × 550 mm for both cope and drag. The pouring system design is fixed. The average wall thickness of the motor housing is 12.7 mm, making it a large thin-walled casting. One casting is produced per mold.
For each kilogram of HT250 iron, the charging materials are: 0.1023 kg pig iron, 0.4249 kg return scrap, 0.5134 kg steel scrap, and 0.0116 kg coke. The carbon emission factors for materials and energy sources used in the calculation are listed below:
| Energy/Material | Carbon Emission Factor |
|---|---|
| Electricity | 0.93 kgCO₂e/kWh |
| Coating | 6.0232 kgCO₂e/kgce |
| Natural gas | 2.1622 kgCO₂e/kgce |
| Alcohol | 0.48 kgCO₂e/kgce |
| Silica sand | 0.02543 kgCO₂e/kgce |
| Pig iron | 2.13 kgCO₂e/kgce |
| Steel | 8.2 kgCO₂e/kgce |
| Tap water | 0.194 kgCO₂e/kgce |
Note: Emissions from fuel use do not include the direct combustion emissions unless explicitly stated. The direct emissions are calculated from chemical reaction stoichiometry.
The melting energy consumption per ton of molten iron varies with pouring temperature. Historical data from the foundry show that a higher pouring temperature leads to higher melting energy consumption. I use a trend line to estimate the melting energy at different pouring temperatures.
5.2 Original Process Plan (Plan 1)
The original process parameters of Plan 1 are shown in the following table:
| Parameter Type | Plan 1 |
|---|---|
| Draft angle | 2° |
| Machining allowance | 9 mm |
| Fillet radius | 8 mm |
| Sprue cross-section size | 70 mm |
| Runner cross-section size | 40/50 × 70 mm |
| Vent cross-section size | Φ110 × 300 mm |
| Pouring temperature | 1446 °C |
| Sand gas evolution | 15.5 mL/g |
| Cooling time | 24 hours |
I verified Plan 1 using ProCAST simulation software. The simulation results indicated that the casting completely solidified in the sand mold, with a few shrinkages located mainly in the gating system and the base of the shell. No critical sand casting defects were found in the casting body. The casting quality was considered acceptable.
The single casting weight is calculated as:
$$M_C = M_P + (V_e + V_a – V_r) \cdot \rho_{HT250}$$
where MP is the net weight of the motor housing, Ve is the volume of machining allowance, Va is the volume of draft angle, Vr is the volume of fillet, and ρHT250 is the density of HT250 (7.2 g/cm³).
For Plan 1, I calculated:
$$V_e = 9 \times [\pi(300.99^2 – 247.65^2) + \pi(410.21^2 – 360.68^2)] = 1,907,008.80 \text{ mm}^3$$
$$V_a = \pi \times 0.5 \times |\tan 2^\circ| \times [(300.99^2 – 247.65^2) \times 300.99 + (410.21^2 – 360.68^2) \times 410.21] = 1,342,319.15 \text{ mm}^3$$
$$V_r = (\pi \times 8^2 – 8^2) \times 1752.98 = 240,266.82 \text{ mm}^3$$
$$M_C = 579.69 + (1,907,008.80 + 1,342,319.15 – 240,266.82) \times 7.2 \times 10^{-6} = 601.36 \text{ kg}$$
The gating system weight is 135 kg. The total resin sand required is 2,747.09 kg.
Using the carbon emission models, I obtained the carbon emissions for Plan 1.
| Carbon Source | Material (kgCO₂) | Energy (kgCO₂) | Undesired (kgCO₂) |
|---|---|---|---|
| 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 |
The total carbon emission for Plan 1 is 2,421.86 kgCO₂. The contributions from IP, CP, and IAP are 78.08%, 19.72%, and 2.20%, respectively. The contributions from material, energy, and undesired carbon sources are 75.82%, 23.92%, and 0.26%, respectively.
5.3 Optimized Process Plan (Plan 2)
I applied the genetic algorithm to minimize the carbon emission function. The ranges for the decision variables are: machining allowance e ∈ [3.5, 9] mm, draft angle α ∈ [0.42°, 3°], fillet radius r ∈ [4, 10] mm, and pouring temperature T ∈ [1440, 1450] °C. These ranges are determined based on the casting design standards and foundry experience to avoid sand casting defects.
The genetic algorithm was implemented in MATLAB 2016b. The convergence curve of the fitness function is shown in Figure 5-8 (as commonly generated by the GA tool). The best individual found is [3.5, 0.42, 10, 1440], meaning machining allowance = 3.5 mm, draft angle = 0.42°, fillet radius = 10 mm, and pouring temperature = 1440 °C. This optimized set constitutes Plan 2.
The optimized process parameters are listed below:
| Parameter Type | Plan 1 | Plan 2 |
|---|---|---|
| Draft angle | 2° | 0.42° |
| Machining allowance | 9 mm | 3.5 mm |
| Fillet radius | 8 mm | 10 mm |
| Sprue cross-section size | 70 mm | 70 mm |
| Runner cross-section size | 40/50 × 70 mm | 40/50 × 70 mm |
| Vent cross-section size | Φ110 × 300 mm | Φ110 × 300 mm |
| Pouring temperature | 1446 °C | 1440 °C |
| Sand gas evolution | 15.5 mL/g | 15.5 mL/g |
| Cooling time | 24 hours | 24 hours |
I verified Plan 2 using ProCAST simulation under the same boundary conditions (filling time 20 s, pouring temperature 1440 °C, air cooling). The simulation showed that the casting completely solidified, and only minor shrinkages were present in the gating system and the base, which are acceptable. No sand casting defects were observed in the main body. The casting quality was confirmed to be satisfactory.
The casting weight for Plan 2 is computed similarly:
$$V_e = 3.5 \times [\pi(300.99^2 – 247.65^2) + \pi(410.21^2 – 360.68^2)] = 741,614.53 \text{ mm}^3$$
$$V_a = \pi \times 0.5 \times |\tan 0.42^\circ| \times [(300.99^2 – 247.65^2) \times 300.99 + (410.21^2 – 360.68^2) \times 410.21] = 281,777.57 \text{ mm}^3$$
$$V_r = (\pi \times 10^2 – 10^2) \times 1752.98 = 375,416.91 \text{ mm}^3$$
$$M_C = 579.69 + (741,614.53 + 281,777.57 – 375,416.91) \times 7.2 \times 10^{-6} = 584.36 \text{ kg}$$
The gating system weight remains 135 kg, and the resin sand required is 2,747.10 kg.
Using the same calculation procedure, the carbon emissions for Plan 2 are:
| Carbon Source | Material (kgCO₂) | Energy (kgCO₂) | Undesired (kgCO₂) |
|---|---|---|---|
| IP | 1709.03 | 142.10 | 0.04 |
| CP | 51.06 | 398.74 | 2.65 |
| IAP | 34.70 | 14.32 | 3.60 |
| Total | 1794.79 | 555.16 | 6.29 |
The total carbon emission for Plan 2 is 2,356.24 kgCO₂. The contributions from IP, CP, and IAP are 78.56%, 19.20%, and 2.23%, respectively. The contributions from material, energy, and undesired carbon sources are 76.17%, 23.56%, and 0.27%, respectively.
5.4 Comparison and Analysis
I compared the carbon emissions of the two plans. The reduction in total carbon emission from Plan 1 to Plan 2 is 65.62 kgCO₂, corresponding to a decrease of 2.71%. This reduction is achieved by optimizing the four process parameters while ensuring that the casting has no sand casting defects.
| Carbon Source | Plan 1 (kgCO₂) | Plan 2 (kgCO₂) | Reduction (kgCO₂) | Reduction (%) |
|---|---|---|---|---|
| Material carbon source | 1836.20 | 1794.79 | 41.40 | 2.25% |
| Energy carbon source | 579.31 | 555.16 | 24.15 | 4.17% |
| Undesired carbon source | 6.35 | 6.29 | 0.06 | 1.00% |
Among the three carbon sources, the energy carbon source shows the largest relative reduction (4.17%), mainly due to the decrease in pouring temperature from 1446 °C to 1440 °C, which lowers the electricity consumption in the melting furnace. The material carbon source shows a moderate reduction of 2.25%, attributed to the reduced machining allowance and draft angle, leading to lower metal and sand consumption. Although the undesired carbon source reduction is small in magnitude, it is still important for environmental protection. The CP-based carbon emission decreases by 5.25%, which is the largest relative reduction among the three parameter categories. The IP-based carbon emission decreases by 39.80 kgCO₂, and the IAP-based emission decreases by 0.76 kgCO₂.
The following table summarizes the changes by process parameter category:
| Parameter Category | Plan 1 (kgCO₂) | Plan 2 (kgCO₂) | Reduction (kgCO₂) | Reduction (%) |
|---|---|---|---|---|
| IP | 1890.96 | 1851.17 | 39.80 | 2.10% |
| CP | 477.52 | 452.46 | 25.06 | 5.25% |
| IAP | 53.38 | 52.61 | 0.76 | 1.43% |
From the analysis, I can draw the following conclusions. First, material carbon emissions account for the largest proportion of total carbon emissions, while energy carbon emissions are the second largest. Second, melting energy consumption is the dominant part of energy carbon emissions, so optimizing the pouring temperature and improving furnace efficiency are crucial for emission reduction. Third, independent process parameters have the greatest impact on material carbon emissions; reducing machining allowance and draft angle effectively lowers material usage. Fourth, to further reduce undesired carbon emissions, more attention should be paid to selecting environmentally friendly materials and treating waste gases. Overall, the proposed method successfully identifies opportunities for carbon reduction in the sand casting process design stage without compromising product quality.
This figure illustrates typical sand casting defects that must be avoided through proper process design. In my optimization, I deliberately constrained the process parameters to ensure that the optimized casting is free from any sand casting defects. The ProCAST simulation results confirmed that both Plan 1 and Plan 2 have only minor shrinkages in the gating system and the base, which do not affect the structural integrity of the final product. Therefore, the carbon emission reduction is achieved without sacrificing casting quality.
6. Conclusions and Future Work
In this thesis, I have presented a systematic framework for calculating and optimizing carbon emissions in sand casting processes based on process parameters. The main contributions of this work are as follows:
- I analyzed the sand casting process and classified process parameters into independent, coupled, and inherent attribute parameters. This classification provides a clear basis for linking process parameters to carbon emissions.
- I established carbon emission models for material, energy, and undesired carbon sources, and integrated them into a unified calculation model. This model can be used at the process design stage to predict carbon emissions without requiring actual production data.
- I formulated a low-carbon optimization problem with machining allowance, draft angle, fillet radius, and pouring temperature as decision variables, and solved it using a genetic algorithm. The optimized parameters successfully reduced carbon emissions.
- I validated the proposed method through a real case study of a motor housing. The optimized process plan achieved a 2.71% reduction in total carbon emissions while maintaining casting quality and avoiding sand casting defects. This reduction is significant for large-scale production.
The results show that material carbon sources are the largest contributor to total carbon emissions, followed by energy carbon sources. The proposed optimization method effectively reduces material and energy carbon emissions by adjusting process parameters. The genetic algorithm proves to be a robust tool for solving the multi-parameter optimization problem in sand casting process design.
This research has some limitations. I did not consider the influence of gating system design and molding process on carbon emissions. In addition, the casting quality was validated only through simulation, not through actual production trials. In the future, I plan to extend the carbon emission model to include the gating system design, such as sprue, runner, ingate, and riser dimensions. I also intend to investigate the relationship between process parameters and specific sand casting defects, allowing the prediction and prevention of defects during the design stage. Furthermore, more comprehensive multi-objective optimization methods that consider both carbon emissions and production costs will be developed.
In summary, the proposed method provides a practical tool for foundry engineers to create low-carbon sand casting processes from the early design stage. By systematically optimizing process parameters, the sand casting industry can move toward sustainable manufacturing with reduced environmental impact while preserving product quality.