As a researcher focused on sustainable manufacturing, I have been deeply involved in studying carbon emissions in industrial processes, particularly in the casting industry. Sand casting is a fundamental manufacturing method for producing metal parts, but it is also a significant source of CO₂ emissions. In many regions, the environmental impact of sand casting parts production is substantially higher compared to more advanced techniques, with energy consumption often double that of similar processes. This has led me to explore methods for reducing carbon footprints through systematic benchmarking. In this paper, I present a comprehensive low-carbon benchmarking model tailored for sand casting parts, leveraging process carbon sources to identify inefficiencies and set quantifiable targets for emission reduction. The goal is to provide a scientific framework that enables foundries to compare their processes against benchmarks and implement targeted improvements. This work is driven by the urgent need to address climate change and promote cleaner production in industries that manufacture sand casting parts.
The concept of energy efficiency benchmarking is widely adopted as a process-oriented approach to enhance energy use by comparing against best practices. However, traditional benchmarking often focuses solely on energy consumption, whereas sand casting involves not only high energy use but also various pollutant emissions across different stages. To achieve true low-carbon manufacturing of sand casting parts, it is essential to adopt a holistic lifecycle perspective that accounts for all carbon emissions, from raw material inputs to final product output. Existing studies have made strides in energy efficiency modeling for casting, but they frequently overlook the nuanced carbon contributions of individual process steps. My approach introduces the notion of “process carbon sources” to break down emissions at the granular level of manufacturing operations. This allows for a more detailed analysis and comparison, enabling foundries to pinpoint specific areas for improvement rather than relying on aggregate metrics. By integrating process carbon sources into a benchmarking model, I aim to bridge this gap and offer a practical tool for reducing the carbon intensity of sand casting parts production.
The core of my model lies in defining and quantifying five fundamental process carbon sources inherent to sand casting operations. These sources represent the basic units of carbon flux emitted during manufacturing activities. First, the Material Consumption Carbon Source (MC) accounts for emissions from consuming various materials, such as sand, binders, and alloys, used in producing sand casting parts. Second, the Undesirable Carbon Source (UC) covers emissions from pollutants and defective parts generated during processes. Third, the Idle Carbon Source (PC) relates to emissions when equipment is in standby or idle mode, waiting to engage in production. Fourth, the Load Carbon Source (LC) captures emissions from equipment under operational load, such as during molding or melting. Finally, the Energy Consumption Carbon Source (EC) includes emissions from non-electric energy sources like fuels. By modeling each sand casting process step with these carbon sources, I can create a detailed emissions profile. For instance, a sand mixing operation might be represented as PC–LC–PC–MC–UC, indicating a sequence of idle, load, idle, material consumption, and undesirable emissions. This structured representation forms the basis for comparing different production routes for sand casting parts.
To operationalize this model, I developed a coding system for sand casting processes, enabling standardized representation and comparison. Based on national standards and industry terminology, I classify processes into four hierarchical levels: major category (e.g., molding department), medium category (e.g., molding segment), minor category (e.g., molding operation), and detailed category. This encoding allows for consistent mapping of manufacturing steps across different foundries producing sand casting parts. For example, a mobile arm sand mixer might be coded as 4122, where 4 represents the sand mixing department, 1 the mixing segment, 2 the resin sand type, and 2 the mobile type. Similarly, a vibrating table compaction process could be coded as 2111. This structured coding facilitates the analysis of process routes and their similarities, which is crucial for effective benchmarking. I also address the complexity of sand casting process structures, which often include serial, parallel, conditional, and repetitive sequences. By applying transformation rules, I convert these mixed structures into standardized serial routes for similarity assessment. This preprocessing step ensures that diverse manufacturing layouts for sand casting parts can be compared on a common basis.
The benchmarking process relies heavily on calculating the similarity between process routes. I define low-carbon benchmarking similarity as a combination of process route similarity and process carbon source similarity. The overall similarity, SP, is given by:
$$ SP = \alpha \cdot S_{DP} + \beta \cdot S_{DC} $$
where \( S_{DP} \) is the process route similarity, \( S_{DC} \) is the process carbon source similarity, and \( \alpha \) and \( \beta \) are weight coefficients determined through expert evaluation. To compute \( S_{DP} \), I consider the hierarchical encoding of processes. For two process routes A and B, the similarity is calculated based on the longest common subsequences at each encoding level. The formula is:
$$ S_{DP} = \frac{\sum_{i=1}^{n} \frac{n+1-i}{n} \sum_{j=1}^{k} L_j(n)}{k \cdot \max\{|A|, |B|\}} $$
Here, \( n \) is the number of encoding levels, \( k \) is the number of sub-routes after transformation, \( L_j(n) \) is the number of identical processes in the \( j \)-th sub-route at level \( n \), and \( \max\{|A|, |B|\} \) is the maximum number of processes in either route. This hierarchical approach ensures that similarities are assessed at appropriate granularities, reflecting the structure of sand casting parts production. For \( S_{DC} \), I compare the counts of the five process carbon sources between routes. The similarity is expressed as:
$$ S_{DC} = \frac{\sum_{i=1}^{N_C} \left(1 – \frac{|C_{A_i} – C_{B_i}|}{\max\{C_{A_i}, C_{B_i}\}}\right)}{N_C} $$
where \( C_{A_i} \) and \( C_{B_i} \) are the counts of the \( i \)-th carbon source type in routes A and B, respectively, and \( N_C \) is the number of carbon source types compared (up to 5). By integrating these similarities, I can identify the most comparable benchmark from a database of historical processes for sand casting parts.
Quantifying carbon emissions is central to my model. I calculate emissions for each process carbon source using established coefficients from standards like PAS2050 and IPCC. For the Idle Carbon Source (PC), emissions depend on equipment idle power and time:
$$ C_{PC} = P_0 \cdot t \cdot E_E $$
where \( P_0 \) is the idle power, \( t \) is the idle time, and \( E_E \) is the carbon emission factor for electricity. For the Load Carbon Source (LC), emissions account for both idle power and additional load:
$$ C_{LC} = (P_0 + \zeta \cdot G \cdot P_w) \cdot t \cdot E_E $$
Here, \( \zeta \) is a unit power loss coefficient, \( G \) is the load weight, and \( P_w \) is the additional power per unit weight. Material Consumption Carbon Source (MC) emissions are derived from the energy used in producing materials:
$$ C_{MC_j} = \sum_{i=1}^{n} \sum_{k=1}^{i} (E_{S_k} \cdot U_{i,j}) \cdot E_E $$
where \( E_{S_k} \) is the electricity consumed in the \( k \)-th processing stage for material \( i \), \( U_{i,j} \) is the amount of material \( i \) consumed in process \( j \), and \( E_E \) is the electricity emission factor. Energy Consumption Carbon Source (EC) emissions are calculated as:
$$ C_{EC_j} = \sum_{i=1}^{n} V_{i,j} \cdot E_i $$
with \( V_{i,j} \) being the amount of energy type \( i \) consumed in process \( j \), and \( E_i \) its emission factor. Undesirable Carbon Source (UC) emissions include waste and pollutants:
$$ C_{UC_j} = \sum_{i=1}^{n} \sum_{k=1}^{i} (E_{S_k} \cdot Q_{i,j} \cdot \phi_i) \cdot E_E $$
where \( Q_{i,j} \) is the quantity of undesirable output \( i \) from process \( j \), and \( \phi_i \) is a treatment difficulty coefficient. These calculations enable precise emission tracking for each step in manufacturing sand casting parts.
To facilitate benchmarking, I organize processes into benchmarking process modules (BPMs), which are groups of related operations. For example, a module might include sand mixing and transportation processes. The carbon emissions of a module are the sum of emissions from its constituent processes. I also define module production capacity (CP_BPM) based on equipment effectiveness, incorporating time utilization, performance rate, and product qualification rate:
$$ CP_{BPM} = \sum_{i=1}^{n} TR_i \cdot DR_i \cdot PR_i = \sum_{i=1}^{n} \frac{N_{q_i}}{t_{a_i} \cdot t_{op_i} \cdot C_{dr_i}} $$
where \( N_{q_i} \) is the number of qualified sand casting parts, \( t_{a_i} \) is total available time, \( t_{op_i} \) is operating time, and \( C_{dr_i} \) is the design capacity. This capacity metric allows for normalized comparisons across different production scales.
The low-carbon benchmarking potential is assessed through three models. First, the process carbon source potential identifies gaps in specific emission types between a target process and a benchmark:
$$ PP_{CS}(L, L_B) = \begin{bmatrix} P_{MC} = MC(L) – MC(L_B) \\ P_{UC} = UC(L) – UC(L_B) \\ P_{PC} = PC(L) – PC(L_B) \\ P_{LC} = LC(L) – LC(L_B) \\ P_{EC} = EC(L) – EC(L_B) \end{bmatrix} $$
Second, the benchmarking process module potential compares emissions per unit capacity for each module:
$$ P_{BPM}(BPM_{L_i}, BPM_{L_{B_i}}) = \frac{C_{L_{BPM_i}}}{CP_{L_{BPM_i}}} – \frac{C_{L_B_{BPM_i}}}{CP_{L_B_{BPM_i}}} $$
Third, the overall unit production capacity carbon emission potential provides a high-level view:
$$ P_C(L) = \frac{C_L}{CP_L} – \frac{C_{L_B}}{CP_{L_B}} $$
These models offer multi-dimensional insights into where and how much improvement is possible for sand casting parts production.
To validate my approach, I applied it to a case study involving a sand casting foundry’s molding production line. The target process route (L1) included steps like mobile sand mixing, conveyor transport, sand filling, vibration compaction, manual molding, spraying, core setting, and box closing. I compared it against four benchmark routes from a database, each with different configurations for producing sand casting parts. Using the encoding system, I represented each route and transformed their structures into serial sequences. For instance, L1 was encoded as 4122-6412-6116-2111-2125-6221-2211-2221-2411. I then calculated process route similarities and process carbon source similarities. The weights \( \alpha \) and \( \beta \) were set to 0.41 and 0.59 based on expert input. The results showed that L1 had the highest similarity with benchmark L3 (SP = 0.8206), making it the optimal benchmark for comparison.
| Target Route | Benchmark Route | Process Route Similarity \( S_{DP} \) | Process Carbon Source Similarity \( S_{DC} \) | Overall Similarity \( SP \) |
|---|---|---|---|---|
| L1 | L2 | 0.472 | 0.6174 | 0.5578 |
| L1 | L3 | 0.667 | 0.9273 | 0.8206 |
| L1 | L4 | 0.556 | 0.7670 | 0.6805 |
| L1 | L5 | 0.500 | 0.2770 | 0.3684 |
Next, I computed carbon emissions using the process carbon source models. I defined three benchmarking modules: BPM1 (mixing and transport), BPM2 (filling, compaction, and mold release), and BPM3 (spraying, core setting, box closing, and pressing). Based on operational data, I constructed carbon emission relationship matrices for both the target and benchmark processes. For example, the matrix for the target process showed emissions in kilograms of CO₂ equivalent (kg CO₂e) across carbon sources and modules:
| Carbon Source | BPM1 | BPM2 | BPM3 |
|---|---|---|---|
| MC | 73.0 | 0 | 21.0 |
| UC | 0.6 | 1.80 | 1.3 |
| PC | 1.8 | 2.10 | 1.3 |
| LC | 4.1 | 2.91 | 2.5 |
| EC | 0 | 0 | 0 |
The benchmark matrix indicated lower emissions, highlighting improvement opportunities. I then calculated production capacities for each module. For instance, BPM1 had a capacity of 0.88 for the target and 0.91 for the benchmark. Using the potential models, I derived specific insights. The process carbon source potential revealed that material consumption (MC) had the largest gap of 30.3 kg CO₂e, suggesting that optimizing material use in sand casting parts production could yield significant reductions. The module potential showed that BPM1 had the highest improvement potential at 32.92 kg CO₂e per unit capacity, indicating that mixing and transport processes are key areas for intervention. Finally, the overall unit capacity potential was 54.98 kg CO₂e per unit capacity, representing the total emission reduction achievable if the target process matched the benchmark’s efficiency. These results provide actionable guidance for the foundry to prioritize节能减排 measures in manufacturing sand casting parts.
In conclusion, the low-carbon benchmarking model based on process carbon sources offers a robust framework for analyzing and improving the environmental performance of sand casting operations. By decomposing emissions into fundamental sources and leveraging similarity calculations, it enables precise identification of benchmarks and quantification of improvement potentials. This approach addresses the limitations of traditional energy-focused benchmarking by encompassing all carbon aspects, from material inputs to waste outputs. For foundries producing sand casting parts, adopting this model can lead to more effective decarbonization strategies, aligning with global sustainability goals. Future work will involve expanding the model to include dynamic factors, such as real-time emission monitoring and adaptive benchmarking, to further enhance its applicability. Additionally, integrating this model with digital twin technologies could enable proactive optimization of sand casting processes, ultimately driving the industry toward greener manufacturing of sand casting parts.