A Novel Collaborative Method to Integrate Carbon Efficiency into Multi-Equipment Operational Coupling for Smart Manufacturing System

Abstract

In the context of carbon neutrality and smart manufacturing, balancing the challenge of carbon and operational efficiency has become a hotspot issue. However, within the specific stage of multi-equipment collaborative manufacturing operational coupling in the production process, multi-state characteristics of equipment operation, multidependencies among operational states, the multi-source of carbon emissions, and spatiotemporal sequence coupling raise the dynamics and complexity of carbon emission modeling and carbon efficiency evaluation. Therefore, a novel methodology to integrate carbon efficiency into a multi-equipment collaboration manufacturing service cell (MECMfg-SC) is proposed in this paper. The stage of multi-equipment collaboration manufacturing operational coupling (MECMfg-OC) in the process of multi-equipment collaboration manufacturing is presented and explained. Then, the operational coupling energy consumption model is constructed based on the MECMfg-OC. The environmental cost performance indicators for smart manufacturing systems, including energy efficiency evaluation (EEe) indicators and carbon efficiency evaluation (CEe) indicators, are proposed. At last, a ball screw smart workshop in a leading Chinese NEV enterprise is introduced to verify the proposed approach. Empirical results confirm the approach’s effectiveness and practical viability.

1. Introduction

The global manufacturing sector faces escalating pressures to reconcile economic growth with environmental sustainability. In 2023, energy-related carbon emissions reached approximately 37.2 Gt, with over 30% attributed to manufacturing [1]. In China, the manufacturing sector accounted for 67.25% of national energy consumption in 2022, contributing more than 11.48 billion metric tons of CO₂ emissions [2]. Recently, the methods for the performance of resource, energy, emission, and economy have attracted more and more attention in academia and enterprises. Researching the green technology innovation of the manufacturing industry from the perspective of carbon neutrality can bring many benefits to the manufacturing industry in implementing carbon emission reduction [3].

Environmental costs are classified into environmental price cost (EPC), representing direct energy consumption, and environmental impact cost (EIC), reflecting emission-related externalities [4]. However, existing models fail to capture their interdependencies or real-time operational dynamics [5]. Ge et al. [6] focus on the multi-feature driven carbon emission time-series coupling model, aiming to reveal the characteristics of carbon emissions and calculate the carbon emissions of laser welding systems. Subsequently, in 2023, Ge et al. [7] conducted integrated decision-making on welding parameters and sequences for multi-characteristic laser welding cells, taking into account both carbon emissions and processing time. They proposed the 6 M characteristics (multi-source, multi-device, multi-state, multi-feature, multi-stage, and multi-sequence) of laser welding cells to reveal the time-series coupling and dynamics of carbon emissions. Although some scholars’ research has revealed the time-series coupling and dynamic characteristics of carbon emissions from manufacturing units, it has not considered overall measurement and assessment from a collaborative perspective. Hierarchical frameworks, such as Sihag’s six-level energy classification [8], and dynamic simulations, such as Alvandi’s economic and environmental VSM (E2VSM) [9], lack granularity in addressing multi-equipment interactions. While life cycle assessment (LCA) [10] and extended exergy analysis (EEA) evaluate sustainability, they remain siloed from operational efficiency metrics.

In the multi-equipment collaboration manufacturing process, multiple production tasks are conducted through information sharing, task allocation, and collaborative work among multi-equipment such as AGVs, industrial robots, and smart machine tools. Furthermore, a specific stage of multi-equipment collaboration manufacturing operational coupling (MECMfg-OC) is ignored in past studies, which refers to multi-equipment performing multi-operation at the same time for a specific task. For example, an industrial robot retrieves materials from an AGV and transfers them to the machine for processing. Upon arriving at the work area, the AGV locks its position and transmits material signals; meanwhile, the robot synchronously identifies, grasps, and transfers the workpiece, and the smart machine responds in real time to complete the collaborative handover. Through spatiotemporal coupling, three pieces of equipment achieve synchronous execution. During the stage of MECMfg-OC, the multi-state of the equipment operation, multi-dependence among states, multi-source of the carbon emission, and spatiotemporal sequence coupling aggravate the complexity of carbon emissions modeling and undermine the accuracy of carbon efficiency evaluation. Therefore, the novel carbon efficiency model and method aiming at MECMfg-OC stage, considering the characteristics of multi-equipment, multi-operation, and time–space coupling, are presented in this study. A specific stage of multi-equipment collaboration manufacturing operational coupling (MECMfg-OC) in the process of manufacturing is presented and explained. Based on this phenomenon, a multi-equipment operational coupling energy consumption model is constructed. Furthermore, environmental costs are categorized into two types, which are environmental price costs and environmental impact costs. An environmental cost model for smart manufacturing systems is then developed, incorporating both the multi-equipment operational coupling effect and the two aforementioned environmental cost categories. Finally, a set of environmental cost performance indicators is proposed, including energy efficiency evaluation (EEe) indicators and carbon efficiency evaluation (CEe) indicators.

The main contributions of this study are summarized as follows:

(1)

Aiming at the lack of research on multi-equipment collaborative operation in smart manufacturing systems, this paper defines and explains the phenomenon of operational coupling, considering the correlation and intensification of intelligent production. Furthermore, the detailed process of smart manufacturing systems is discussed from the aspects of spatiotemporal constraints, operational coupling, and collaborative modes. Achieving an accurate description of smart manufacturing systems from the micro level.

(2)

Constructing an energy consumption cost model for smart manufacturing systems, which breaks through the traditional manufacturing system energy consumption method based on manufacturing units. And can more comprehensively and accurately calculate the energy consumption of smart manufacturing systems at the micro level.

(3)

A manufacturing system energy consumption evaluation model including EPC and EIC is proposed based on the material–energy flow coupling model (MEFCM) and extended value stream mapping (EVSM). This model includes more information than the traditional VSM model and can better discover the problems existing in the manufacturing system, thus pointing out more accurate directions for the improvement of the manufacturing system.

The rest of this paper proceeds as follows. Section 2 reviews the current status of coupling effects in smart manufacturing, the environmental cost identification and quantification, assessment indicators and methods, and energy value mapping studies. Section 3 elaborates on the methodology of integrating carbon efficiency into the manufacturing process in detail. In Section 4, the method is verified by a famous new energy vehicle enterprise case. Conclusions and future studies are outlined in the Section 6.

2. Literature Review

2.1. Coupling Effects in Smart Manufacturing

The modern manufacturing system is a typical complex system of multi-equipment, multi-states, and multi-units, and the core function is to produce high-quality products that meet task requirements continuously and efficiently [11,12,13]. To accommodate growing customization demands, traditional mass production has transitioned toward flexible smart factories [14]. Such factories typically integrate multiple intelligent manufacturing units [15], and agile systems coupling smart machine tools, industrial robots, and AGVs into interdependent material processing–handling [16].

The manufacturing cell concept originates from cellular manufacturing. Ostrosi and Fougères [17] developed an intelligent virtual manufacturing cell formation method for configuration-based design and manufacturing (CBDM). Deliktas et al. [18] and Derya et al. [19] accounted for flexible process routings, inter-cell transportation, and sequence-dependent setups in flexible job-shop cellular manufacturing. Zhou et al. [20] established a knowledge-driven digital twin manufacturing cell (KDTMC) framework enabling intelligent manufacturing through autonomous perception, simulation, and control. Tian et al. [21] systematically reviewed multi-criteria decision-making (MCDM) applications in green logistics, with their insights advancing reconfiguration solutions for complex cellular systems. Iqbal and Al-Ghamdi [22] resolved energy-efficient multi-objective flow-line scheduling with sequence-dependent constraints. Hong et al. [23] extended this work by incorporating eligibility constraints.

Driven by demands for enhanced flexibility, efficiency, and responsiveness in manufacturing, cellular production systems continuously balance stability and adaptability. This dynamic necessitates replacing physical demarcation of cell boundaries with process-driven logical units, rendering traditional static energy models inadequate. Beyond baseline machine operation energy, cellular manufacturing energy consumption must also fulfill intended manufacturing functions [24]. Contemporary flexible and intelligent manufacturing environments exhibit multi-dimensional coupling effects during system operation, significantly complicating holistic energy assessment.

2.2. Energy Consumption and Carbon Emissions Accounting for Manufacturing Processes

As an important industrial support for the economies of various countries, the manufacturing industry will need to transform and upgrade to realize the vision of carbon neutrality [25,26,27]. Current trends in manufacturing are moving towards the adoption and implementation of smart manufacturing. While aligning with global economic development trends, smart manufacturing also promotes the green and digital transformation of the manufacturing industry [28]. In environmental price cost accounting, Gu et al. [29] developed an energy consumption model that accounts for full states in machining processes. Their approach links workpiece design features, machine tool parameters, and process parameters to energy consumption. Sihag and Sangwan [5] established a six-tier energy classification framework. Lv et al. [30] characterized processing energy consumption using a multi-energy source and dynamic mechanism analytical model for machine tools. A review by Yusuf et al. [31] synthesizes energy consumption patterns and minimization strategies for machine tools across manufacturing processes. Alghieth [32] presents a multi-modal deep learning framework that integrates Convolutional Neural Networks (CNNs) for defect detection, Recurrent Neural Networks (RNNs) for predictive energy consumption modeling, and Reinforcement Learning (RL) for dynamic energy optimization to enhance industrial sustainability.

For carbon accounting, Zhou et al. [33] investigated the correlation between processing carbon emissions and cutting power, establishing a quantitative carbon emission model for machining processes via experimental fitting. Shang et al. [34] proposed an equivalent carbon emission quantification method for the turning process, incorporating information flow, energy consumption, resource loss, and waste generation. Some scholars are extending research towards dynamic carbon emissions and optimization. Yao et al. [35] proposed a framework for quantifying carbon emissions in four-layer machining using IoT and material and energy flow analysis (MEFA) technology, validating it with a case study. Ge et al. [36] introduced the concept of a “meta carbon emission block” consisting of static and variable blocks, suggesting a data-driven accounting method for manufacturing systems. Liu et al. [37] analyzed carbon emission characteristics in directed energy deposition and built an optimization model for the process.

Overall, existing research primarily employs three approaches: system decomposition modeling, mechanistic analysis, and experimental or statistical analysis. Despite these advances, significant limitations remain in supporting dynamic coupled environmental price-cost accounting and carbon efficiency optimization.

2.3. Lean Manufacturing and Life Cycle Assessment

Originating from the Toyota Production System, which can be traced back to the 1950s [38], its core focus lies in cost reduction through the elimination of non-value-added activities [39]. Over the past two decades, the widespread application of lean tools and techniques has significantly reduced waste across contexts, from shop-floor production to cross-functional enterprise processes [40]. Raoufi and Haapala [41] aim to facilitate sustainability performance analysis in manufacturing processes and systems via unit manufacturing process (UMP) modeling, implemented within an accessible public platform for product design and manufacturing analysis. Furthermore, integrating methodologies like VSM with sustainability is receiving significant attention. Araujo Galvao et al. [42] focus on enhancing traditional time-based VSM by adding sustainability indicators related to materials, environment [43], energy, water, emissions, transport, waste, efficiency (OEE), physical work index, noise level, and life cycle assessment [44]. Huang et al. [45] evaluated total-factor energy productivity based on the stochastic frontier analysis (SFA) approach and then analyzed the energy productivity effects of industrial intelligence. Some studies have investigated the impacts of environmental regulation policies on the performance of smart manufacturing enterprises, promoting the green development of the smart manufacturing industry in the context of Industry 4.0, and fostering the synergistic development of smart manufacturing with carbon neutrality [46].

Life cycle assessment (LCA) has grown into a widely used method for addressing the environmental aspects of products and services [47]. Piron et al. [48] address the gap by introducing innovative approaches to life cycle assessment through Industry 4.0 technologies. Increasing data availability and rapid development of machine learning (ML) approaches offer new opportunities to advance LCA [49]. Ziyadi and Al-Qadi [50] emphasize that objective uncertainty quantification (UQ) in product LCA is critical for decision-making, with environmental impacts measurable either directly or via modeling; Akrami et al. [51] document a recent surge in studies applying AI (notably ML) to environmental sustainability and climate change, with growing interest in combining ML and LCA to expand the breadth and depth of LCA research. Considering environmental impacts across the full life cycle of machine tools, Liang et al. [52] established a comprehensive sustainability assessment index encompassing energy consumption, processing costs, waste management, environmental impact, personal health, and safety, with each component modeled.

2.4. Summary

Research on energy consumption and environmental cost of smart manufacturing systems has made remarkable progress. However, three key research challenges remain: (1) Lack of systematic research and in-depth analysis of the operation coupling effect of intelligent manufacturing units. (2) The environmental cost model of smart manufacturing systems is not developed enough, and existing accounting methods struggle to adapt to modern multi-equipment collaborative manufacturing scenarios. (3) Carbon efficiency assessment and operational optimization of multi-equipment coupling still face challenges.

3. Methodology

In order to address the challenge of integrating carbon efficiency into multi-equipment operational coupling for smart manufacturing systems, the multi-equipment operational coupling effect is proposed and explained in this study. A novel method integrating carbon efficiency into the multi-equipment collaboration manufacturing service cell (MECMfg-SC) is proposed in this section. The method consists of three steps, as shown in Figure 1.

Step 1: The stage of multi-equipment collaboration manufacturing operational coupling (MECMfg-OC) is proposed and defined. It establishes a model foundation for carbon efficiency optimization. Then, the state EnergyBlock (SEB) and component EnergyBlock (CEB) are applied to characterize the energy consumption behaviors of each equipment in various operating states for MECMfg-SC.

Step 2: Environmental cost modeling for smart manufacturing systems tackles incomplete cost quantification. First, identify the environmental cost assessment boundary of the smart manufacturing system. Second, the environmental cost model of the smart manufacturing system is constructed, considering the multi-equipment operational coupling and the two environmental cost classes: the EPC model and the EIC model.

Step 3: Energy efficiency and carbon efficiency indicators are proposed. These indicators are integrated into value stream mapping. The mapping simultaneously visualizes value streams, carbon emission streams, and energy consumption streams. This integration exposes system bottlenecks and provides decision support for continuous optimization.

3.1. Multi-Equipment Operational Coupling Energy Consumption Modeling

In the smart manufacturing system, multi-equipment collaboration manufacturing refers to the process of completing production tasks together through information sharing, task allocation, and collaborative work between equipment such as AGV, industrial robots, and smart machine tools [53]. In collaborative manufacturing, the time–space coupling phenomenon of equipment operation will inevitably occur due to the strong coupling characteristics of multi-equipment timing dependencies. This phenomenon is referred to as MECMfg-OC in this study. Because this phenomenon has not been explained in past studies, the collaborative microprocess and the energy coupling characteristics of MECMfg-OC are discussed in this section.

3.1.1. Presentation and Explanation of the Stage for MECMfg-OC

MECMfg-OC refers to a multi-dimensional interdependent relationship formed by heterogeneous equipment clusters $E = {\left\{{eq}_i\right\}}_{i = 1}^n$ under manufacturing task T. This relationship arises from dynamic interactions of information flow (status data, control commands), material flow (workpiece transfer), and energy flow (power coordination).

Essentially, it is a co-evolution process between the system state space $S = {⨂}_{i = 1}^n{S}_i$ and the collaborative operation set O constrained by tasks. This process is formally defined as a nonlinear mapping:

$$\mathsf{\Gamma }:\left(T,C,S\right)\mapsto \left(S^{’},O\right)$$

(1)

where $T$ is the manufacturing task, $T\subseteq A^{*}$, $A$ is the partial order set of meta-operation sequences; $C$ is the dynamic coupling matrix, quantifying the degree of coupling between equipment; $C\in {\mathbb{R}}^{n \times n}$ is the system state space; $S = {⨂}_{i = 1}^n{S}_i$, $S_i$ is the local state of the equipment $d_i$. The symbol $⨂$ represents the tensor product (also known as the Kronecker product) operator. In the system state space $S = {⨂}_{i = 1}^n{S}_i$, it is used to describe the combined state space of multiple heterogeneous devices in the multi-device collaborative manufacturing system.

(1)

The stage of multi-equipment collaboration manufacturing operational coupling

MECMfg-OC has various manifestations and diverse structures. To explain the stage of MECMfg-OC more precisely, this paper proposes a novel production execution logic model for multi-equipment collaborative manufacturing service cells (MECMfg-SC). Developed through a conceptual extension of the SE and the MNOSE, the model builds upon the research of Jiang et al. [54] and Liu et al. [55]. The MECMfg-OC is described through the representation and integration of FICM (flows of information, control, and material).

$$\mathrm{MECMfg-SC} = 〈D,T,C,F_{FICM},S,O〉$$

(2)

where $F_{FICM} = F_I⨁F_M⨁F_{\epsilon }$ is information, material, and energy three-flow fusion operator. The fundamental elements of SE and MNOSE are depicted in the lower dashed box of Figure 2, with detailed definitions provided by Su et al. [56].

The stage of MECMfg-OC can be classified into three types of operational coupling service cells (OC-SC), as shown in Figure 3:

(a)

TEP operational coupling (Transport–Executor–Processor operational coupling): an equipment collaboration mechanism driven by material transfer as the core to realize inter-processor and transporter material flow through dynamic path planning of executors.

(b)

PET operational coupling (Processor–Executor–Transport operational coupling): the process is the core of the reverse material feedback collaboration mechanism, emphasizing the processor state of the transmission path of dynamic adjustment.

(c)

PEP operational coupling (Processor–Executor–Processor operational coupling): a multi-processor synergy mechanism for complex process chains, realizing the transfer of process parameters through the intelligent intermediary of executors.

(2)

AGV–robot–machine tool collaborative manufacturing chain

Furthermore, more complex MECMfg-SCs can be constructed by combining multiple OC-SCs. For example, an industrial robot retrieves materials from an AGV and transfers them to an intelligent machine tool for processing. As shown in Figure 4, this forms an AGV–robot–machine tool collaborative manufacturing chain. Such chains can be effectively described through the composition of multiple OC-SCs.

In the AGV–robot–machine tool collaborative manufacturing chain:

(a)

The AGV trolley $a_1$ responsible for transporting the parts to be processed sends a collaborative start signal (CSS) to the industrial robot $r_1$ at $t_1$ and starts to perform deceleration operations until $t_2$ decelerates to the specified position. After receiving the CSS, $r_1$ performs the slowdown, which is completed at $t_2$, and then sends the collaborative completion signal (CCS) to $a_1$, and then performs the transfer. After $a_1$ receives the signal, it proceeds to the next operation.

(b)

At $t_3$, $r_1$ sends a CSS to the CNC machine $m_1$ and continues to move towards it. After receiving the CSS, $m_1$ performs the ready for processing. At $t_4$, the industrial robot completes the unloading and places the parts to be processed on the $m_1$ table. Subsequently, the CNC machine tool performs the processing and transmits a CCS to the industrial robot $r_1$, which performs a reset operation.

Similarly, at $t_6$, processing completion on the CNC machine triggers its idle state transition, concurrently issuing a CSS to the industrial robot. CSS receipt initiates robotic loading operations, finalized at $t_7$ with a CCS return to the CNC machine. Upon CCS reception at $t_8$, the CNC machine enters standby mode. Simultaneously, the robot commences workpiece transfer and dispatches CSS to the AGV at $t_9$. The AGV responds by decelerating to approach its target position, achieving alignment at $t_{10}$—synchronizing with robotic unloading completion. Post-unloading, the robot transmits CCS to the AGV, activating loaded material transportation.

3.1.2. Energy Consumption Quantification for Multi-Equipment Operational Coupling

Machine energy consumption depends primarily on operational states (standby, idle, processing, etc.), with state transitions driven by functional component activation or deactivation [57]. As depicted in Figure 5, the MECMfg-SC power profile characterizes equipment state transitions and power variations. The profile, after segmentation and linearization, transforms into a series of state EnergyBlocks superimposed with component EnergyBlocks. Characteristic profile inflection points mark state transitions, defining SEB boundaries. The energy consumption characteristic function is given in Equation (3).

$$p_i = \left\{\begin{array}{c}{f}_{i,1}\left(t\right), \forall t\in \left[t_0,t_1\right], \\ f_{i,2}\left(t\right), \forall t\in \left[t_1,t_2\right], \\ \cdots \\ f_{i,r}\left(t\right), \forall t\in \left[t_{r-1},t_r\right], \end{array}\right.\left.\begin{array}{c}operation state 1\\ operation state 1\\ \cdots \\ operation state 1\end{array}\right\}$$

(3)

where $p_i$is the energy consumption characteristic function of the $i$-th device. $f_{i,r}\left(t\right)$ is the energy consumption characteristic function of the $i$-th device in the $r$-th operation state.

The state EnergyBlocks are categorized into static state EnergyBlock (SSEB) and variable state EnergyBlock (VSEB).

According to the concept of state EnergyBlock, as shown in Figure 4, the energy consumption accounting approach for the OC-SC is presented. The details are as follows:

$${SEB}^I = P^i \times S^i \times T^i = \left[{SEB}_1^i,\cdots ,{SEB}_r^i,\cdots {SEB}_R^i\right]$$

(4)

$$P^i = \left[P_1^i,\cdots ,P_k^i,\cdots P_K^i\right]$$

(5)

$$S^i = \left[\begin{array}{c}{S}_{\mathrm{1,1}}^i,\cdots ,S_{1,r}^i,\cdots S_{1,R}^i\\ ⋮ \cdots ⋮ \cdots ⋮ \\ S_{k,1}^i,\cdots ,S_{k,r}^i,\cdots S_{k,R}^i\\ ⋮ \cdots ⋮ \cdots ⋮ \\ S_{K,1}^i,\cdots ,S_{K,r}^i,\cdots S_{K,R}^i\end{array}\right]$$

(6)

$$T^i = \left[\begin{array}{c}{T}_{\mathrm{1,1}}^i,\cdots ,T_{1,r}^i,\cdots T_{1,R}^i\\ ⋮ \cdots ⋮ \cdots ⋮ \\ T_{k,1}^i,\cdots ,T_{k,r}^i,\cdots T_{k,R}^i\\ ⋮ \cdots ⋮ \cdots ⋮ \\ T_{K,1}^i,\cdots ,T_{K,r}^i,\cdots T_{K,R}^i\end{array}\right]$$

(7)

where ${SEB}^i$, $P^i$, $S^i$, and $T^i$ denote the energy consumption matrix for the operation state by the $i\mathrm{t}\mathrm{h}\left(i = \mathrm{1,2},3,\cdots ,n\right)$ device, component power matrix, component activation state matrix, and state duration matrix, respectively; $P_k^i$ is the power level of the $k$th $\left(k = \mathrm{1,2},3,\cdots ,K\right)$ component; $S_{k,r}^i$ denotes the activation state of the $k$th component under the $r$th $\left(r = \mathrm{1,2},3,\cdots ,R\right)$ operation state. If the component is activated $S_{k,r}^i$ equals to 1 for 0-1 distribution model, otherwise 0; if the component is activated, $S_{k,r}^i$ is a constant greater than 1 for the discrete distribution model, otherwise $S_{k,r }^i$ = 0. $T_{k,r}^i$ denotes the duration of the $r$th operation state triggered by energy consumption.

Equipment energy consumption depends not only on its state but also on the state of other equipment in the collaborative manufacturing network, considering the dynamic characteristics of multi-equipment collaborative manufacturing operations. To accurately quantify this coupling effect, a state energy consumption coupling model is proposed. Therefore, the MECMfg-OC energy consumption model is given in Equation (8):

$${EC}_{coup} = {\sum }_{i = 1}^n{\sum }_{r = 1}^m{SEB}_r^i$$

(8)

where ${SEB}_r^i$ denotes the $r$th operation state of the $i$th device.

3.2. Environmental Cost Modeling for Smart Manufacturing System

In this section, the environmental cost model for the smart manufacturing system is constructed and considers the multi-equipment operational coupling.

3.2.1. Identifying the Environmental Cost

The environmental cost of smart manufacturing systems consists of two components: environmental price cost and environmental impact cost. As shown in Figure 6, each manufacturing process carries a corresponding environmental price cost. The energy consumption in manufacturing processes inevitably produces environmental impact costs [58].

(1)

Environmental price costs

The environmental price cost refers to the direct or indirect energy costs of equipment (e.g., machine tools, AGVs, industrial robots) [59] in manufacturing systems. Energy consumption constitutes the fundamental source of carbon emissions in manufacturing systems.

(2)

Environmental impact costs

The environmental impact cost pertains to carbon emissions released during manufacturing processes. As shown in Figure 6, the boundary of environmental impact costs for the smart manufacturing system where carbon emissions derived from complex sources, including material removal, equipment degradation, energy consumption, and waste generation.

3.2.2. Environmental Cost Model

Smart manufacturing systems consist of multiple manufacturing units (such as MECMfg-SC and assembly units), transportation systems between manufacturing units, and auxiliary and smart service units. The environmental cost model of the intelligent workshop is constructed, considering the MECMfg-OC and the two environmental cost classes: EPC and EIC models. The following are the specific formulae for the calculation:

$${EPC}_{IMS} = {EPC}_{M-Pro} + {EPC}_{OC-SC} + {EPC}_{tran} + {EPC}_{Ind}$$

(9)

where ${EPC}_{IMS}$ is environmental price costs of smart manufacturing system; ${EPC}_{M-Pro}$ is environmental price cost of MECMfg-SC processing stage; ${EPC}_{OC-SC}$ is environmental price cost of MECMfg-SC operational coupling stage; ${EPC}_{tran}$ is environmental price cost of transportation system between manufacturing units; ${EPC}_{Ind}$ is environmental price cost of auxiliary and intelligent services units.

$${EIC}_{IMS} = {EIC}_{M-Pro} + {EIC}_{OC-SC} + {EIC}_{tran} + {EIC}_{Ind}$$

(10)

where ${EIC}_{IMS}$ is environmental impact costs of smart manufacturing system; ${EIC}_{M-Pro}$ is environmental impact cost of MECMfg-SC processing stage; ${EIC}_{OC-SC}$ is environmental impact cost of MECMfg-SC operational coupling stage; ${EIC}_{tran}$ is environmental impact cost of transportation system between manufacturing units; ${EIC}_{Ind}$ is environmental impact cost of auxiliary and intelligent services units.

Next, the environmental price cost and environmental impact cost calculations for smart manufacturing systems consisting of MECMfg-SC will be discussed in detail.

(1)

Environmental price costs

To more accurately calculate the environmental cost of MECMfg-SC, as shown in Figure 4 and Figure 5, this section further divides the MECMfg-SC into two sub-stages, namely the processing stage and the operation coupling stage.

(a)

Processing stage of MECMfg-SC

In MECMfg-SC, smart machine tools as the primary energy-consuming equipment during processing. Electrical energy is converted into mechanical energy through servo motors and auxiliary components by these tools. Based on machine tool operating states, energy consumption is categorized into standby, air-cutting, and cutting energy consumption [60]. The environmental price cost for this stage is calculated as follows:

$${EPC}_{M-Pro} = R_{elec}·{EC}_{M-Pro} = R_{elec}·{\sum }_{r = 1}^m{SEB}_r^{machine}$$

(11)

where $R_{elec}$ is electricity price; ${EC}_{M-Pro}$ is energy consumption at the processing stage of MECMfg-SC; ${SEB}_r^{machine}$ is state EnergyBlocks for machine tools and $r =$(standby, air-cutting, cutting).

The processing stage of MECMfg-SC generates carbon emissions primarily originating from equipment energy consumption, material consumption, and waste treatment, categorized as direct and indirect emissions. The carbon emissions generated in the processing stage of MECMfg-SC, it is mainly composed of direct carbon emissions ${CE}_{M-Pr{o}_{dir}}$ and indirect carbon emissions ${CE}_{M-Pr{o}_{indir}}$ generated by equipment energy consumption, material consumption, and waste treatment. Therefore, the environmental impact cost of the processing stage is calculated as follows:

$${{EIC}_{M-Pro} = CE}_{M-Pr{o}_{dir}} + {CE}_{M-Pr{o}_{indir}}$$

(12)

$${CE}_{M-Pr{o}_{indir}} = {CE}_{ec} + {CE}_{mac} + {CE}_{wat}$$

(13)

where ${CE}_{ec}$ is energy consumption carbon emissions; ${CE}_{mac}$ is material consumption carbon emissions; ${CE}_{wat}$ is waste treatment carbon emissions.

Among them, ${CE}_{M-Pr{o}_{dir}}$ refers to carbon emissions generated by chemical reactions during processing, such as fuel combustion, material heat decomposition, etc. ${CE}_{ec}$ mainly considers the carbon emission generated by power consumption in the production process. ${CE}_{mac}$ refers to the carbon emissions generated during the preparation process of raw materials, auxiliary materials, and primary/secondary energy (except electric energy). ${CE}_{wat}$ considers the carbon emissions generated during the treatment of waste debris, waste liquid, waste sand, and other substances. Consequently, their detailed calculations are as follows:

$${CE}_{ec} = {EC}_{M-Pro}·{CEF}_{elec}$$

(14)

$${CE}_{mac} = {CE}_{rm} + {CE}_{ma}$$

(15)

$${CE}_{wat} = {CE}_{cool,w} + {CE}_{chip,w}$$

(16)

where ${CEF}_{elec}$ is electric energy–carbon emission factor; ${CE}_{rm}$ is raw material consumption carbon emissions; ${CE}_{am}$ is auxiliary material consumption carbon emissions; ${CE}_{cool,w}$ is waste-cutting fluid/waste lubricating oil treatment carbon emissions; ${CE}_{chip}$ is waste chips treatment carbon emissions.

(b)

Operational coupling stage of MECMfg-SC

In Section 3.1.2, the known operational coupling energy consumption ${EC}_{OC-SC}$ is calculated as shown in Equation (14). The environmental impact cost for this stage is calculated as follows:

$${EC}_{OC-SC} = {\sum }_{i = 1}^n{\sum }_{r = 1}^m{SEB}_r^i$$

(17)

$${EPC}_{OC-SC} = R_{elec}·{EC}_{OC-SC}$$

(18)

$${EIC}_{OC-SC} = {EC}_{OC-SC}·{CEF}_{elec}$$

(19)

where ${EC}_{OC-SC}$ is energy consumption of operational coupling.

(2)

Environmental cost model of transportation between manufacturing units

In a smart manufacturing system, for a product manufacturing chain consisting of multiple manufacturing units, the environmental cost includes not only the environmental cost of the manufacturing units but also the environmental cost of transportation between the units. The transportation equipment between manufacturing units in the smart manufacturing system includes AGVs, conveyor lines, and so on. The calculation of the environmental cost of transportation between manufacturing units is as follows:

$${EPC}_{tran} = R_{elec}·{EC}_{tran}$$

(20)

$${EC}_{tran} = \sum p_{tran}·t_{tran}$$

(21)

$${EIC}_{tran} = {EC}_{tran}·{CEF}_{elec}$$

(22)

where ${EC}_{tran}$ is energy consumption of transportation between manufacturing units; $P_{tran}$ is power for transportation equipment; $t_{tran}$ is transportation equipment working hours.

(3)

Environmental costs model of factory facilities and smart service units

Auxiliary equipment for maintaining the production environment in the smart manufacturing system includes exhaust systems, lighting systems, heating, air conditioning, and other auxiliary production facilities. Equipment that provides smart service mainly includes smart integrated consoles, smart terminals, and other smart service equipment. Data on the energy loss of these power-consuming public infrastructures over a period of time can be obtained by measuring the power, counting the time of use, and finally calculating it. The environmental costs of factory facilities and smart service units are calculated as follows:

$${EPC}_{Ind} = R_{elec}{\sum }_{b = 1}^{\xi }{\sum }_{i = 1}^n\left(P_{\xi }^{i,a} \times T_{\xi }^{i,a}\right)$$

(23)

$${EIC}_{Ind} = {CEF}_{elec}{\sum }_{b = 1}^{\xi }{\sum }_{i = 1}^n\left(P_{\xi }^{i,a} \times T_{\xi }^{i,a}\right)$$

(24)

where $P_{\xi }^{i,a}$ is power for factory facilities and smart service equipment; $t_{\xi }^{i,a}$ is working hours for factory facilities and smart service equipment; $b = 1~\xi$ are the exhaust equipment, lighting equipment, heating equipment, air conditioning, intelligent service equipment, and other auxiliary equipment.

3.3. Environmental Cost Performance Evaluation for MECMfg-SC

The complexity of manufacturing systems contains many productivity variables that affect energy consumption and carbon emissions, making environmental cost diagnosis a challenging task. In this section, a set of environmental cost performance indicators for MECMfg-SC is proposed. Then, the environmental and carbon-oriented Value Stream Map (EC-VSM) is implemented to integrate the environmental cost performance indicators into VSM to visualize the value stream, carbon emission stream, and energy consumption pattern of MECMfg-SC to support production decisions.

3.3.1. Energy Efficiency Evaluation Indicators

EEe (energy efficiency evaluation indicators) is inspired by OEE; this includes production-oriented energy performance indicators [61] and MECMfg-SC-oriented energy efficiency indicators. The production-oriented energy performance indicators are defined below.

• Theoretical Energy Consumption (TEC): The amount of energy required to complete a specified operation under ideal conditions, based on theoretical calculations of chemical or physical laws.
• Direct Energy Consumption (DEC): The energy consumption in the production process that is directly used in core operations such as processing and handling, including value-added energy consumption and non-value-added energy consumption.
• Value-added Energy Consumption (VEC): Energy consumption in direct energy consumption for activities that directly enhance the value of the product, such as processing, assembly, effective testing, and other customer-recognized value-creation processes.
• Non-value-added Energy Consumption (NVEC): The portion of direct energy consumption that does not directly create product value, which is categorized into two categories: (a) essential non-value-added energy consumption, e.g., equipment preheating, safety inspection; and (b) purely wasteful energy consumption, e.g., idle equipment, waiting for empty loads, inefficient operation.
• Overall Energy Consumption (OEC): The actual energy consumption to produce a given number of qualifying parts, including direct energy consumption and indirect energy consumption (e.g., energy consumption of non-core operations such as plant facilities, intelligent service equipment, etc., that maintain the production environment). The calculation of OEC takes into account all energy consumption within the process boundary.

Building upon this foundation, energy efficiency indicators for smart manufacturing systems and MECMfg-SC are established. The specific definitions are as follows.

(1)

Manufacturing system

• Energy Utilization Rate ${ER}_S^i$, ratio of VEC to OEC, is shown in Equation (25). The percentage of total energy consumption of a manufacturing system that is used to directly create product value. The higher the value, the less energy is wasted.

$${ER}_S^i = {\displaystyle \frac{\sum _i^n\sum _{r_p}{SEB}_{r_p}^{i_{processor}} + \sum _i^n\sum _{r_p}{SEB}_{r_e}^{i_{executor}} + \sum _i^n\sum _{r_t}{SEB}_{r_t}^{i_{transport}}}{{EC}_{M-Pro} + {EC}_{OC-SC} + {EC}_{tran} + {EC}_{Ind}}}$$

(25)

where $r_p =$ cutting, $r_e =$ transferring, $r_t =$ load transportation.

(2)

Manufacturing unit

• Value-Added Energy Consumption Ratio ${ER}_M^i$, ratio of VEC to DEC, is shown in Equation (26). The effectiveness of the direct energy consumption of a manufacturing cell for value creation. The higher the value, the less non-value-added energy is consumed (e.g., idle, waiting).

$${ER}_M^i = {\displaystyle \frac{\sum _i^n\sum _{r_p}{SEB}_{r_p}^{i_{processor}} + \sum _i^n\sum _{r_p}{SEB}_{r_e}^{i_{executor}}}{{EC}_{M-Pro} + {EC}_{OC-SC}}}$$

(26)

• Operational Coupling Energy Consumption Ratio ${ER}_M^{OC}$, ratio of ${EC}_{OC-SC}$ to ${EC}_{IMS}$, is shown in Equation (27). The lower the value, the better the equipment collaboration performance.

$${ER}_M^{OC} = {\displaystyle \frac{{EC}_{OC-SC}}{{EC}_{M-Pro} + {EC}_{OC-SC}}}$$

(27)

3.3.2. Carbon Efficiency Evaluation Indicators

In order to comprehend and analyze the carbon emission characteristics of manufacturing systems and manufacturing units from multiple perspectives, two carbon efficiency indicators were defined based on parameters that measure the effectiveness of functional services, such as the economic return rate of manufacturing systems and the production rate of manufacturing units. These indicators are the economic return rate carbon efficiency (ERRC) for manufacturing systems and the production rate carbon efficiency (PRC) for manufacturing units. The specific definitions are as follows.

(1)

Manufacturing system

The parameter ${\eta }_b$, referred to as the economic return rate of carbon efficiency ($\mathrm{C}\mathrm{N}\mathrm{Y}/\mathrm{k}\mathrm{g}{\mathrm{C}\mathrm{O}}_2$), represents the ratio between the economic return of the manufacturing system and its average carbon emissions per unit time. This indicator is macroeconomically oriented and reflects the economic value generated per unit of carbon emission by the manufacturing system. A higher ${\eta }_b$ value indicates that the system achieves greater economic returns with lower associated carbon emissions, signifying better overall performance. Conversely, a lower ${\eta }_b$ value suggests reduced efficiency and environmental sustainability of the manufacturing system.

$${\eta }_b = b/{CE}_T^s$$

(28)

where $b$ is the economic return rate of the manufacturing system ($\mathrm{C}\mathrm{N}\mathrm{Y}/\mathrm{h}$), ${CE}_T^s$ is average carbon emissions per unit time of manufacturing system ($\left(\mathrm{k}\mathrm{g}{\mathrm{C}\mathrm{O}}_2\right)/\mathrm{h}$).

(2)

Manufacturing unit

${\eta }_p$ is production rate carbon efficiency ($\mathrm{p}\mathrm{i}\mathrm{e}\mathrm{c}\mathrm{e}/\mathrm{k}\mathrm{g}{\mathrm{C}\mathrm{O}}_2$), which is the ratio of the production rate of a manufacturing unit to the average carbon emissions per unit of time. It measures the processing performance of a manufacturing unit. The higher the value, the lower the carbon emissions generated by the manufacturing unit’s processing technology and the better the processing performance of the manufacturing unit; conversely, the lower the value, the worse the performance.

$${\eta }_p = p/{CE}_T^s$$

(29)

where $p$ is the production rate of a manufacturing unit ($\mathrm{p}\mathrm{i}\mathrm{e}\mathrm{c}\mathrm{e}/\mathrm{h}$), ${CE}_T^M$ is average carbon emissions per unit of time ($\mathrm{k}\mathrm{g}{\mathrm{C}\mathrm{O}}_2/\mathrm{h}$).

3.3.3. MECMfg-SC State Description Based on the EC-VSM

Value stream mapping, also referred to as material and information flow mapping, is a lean manufacturing tool employed to analyze, assess, and optimize specific work processes within production systems [62]. It facilitates the visualization of current process flows, thereby enabling practitioners to identify potential areas for improvement. As illustrated in Figure 7, an EC-VSM is applied to represent the value flow, carbon emission pathways, and energy consumption patterns across the production process.

Under each process box, a data table is added to present key metrics, such as cycle time (C/T). Beyond productivity-related data, metrics about energy attributes must also be included. One typical example is the non-processing load factor (NPLF), which is defined as the ratio of the power consumption level during non-processing states (e.g., standby mode) to that during active processing states. Notably, a higher NPLF indicates greater energy-saving potential, which can be achieved through engineering interventions (e.g., shutting down unnecessary components). Additionally, the aggregate lean energy indicator (see Section 3.3) can be incorporated into these tables.

Beneath each data table, a time and energy bar chart is included to visualize the room for improvement in each process. Specifically, the purple bar represents value-added time/energy, while the gray bar denotes non-value-added time/energy.

4. Case Study

The ball screw pair is a critical precision transmission component widely applied in new energy vehicles, particularly in systems such as electric power steering (EPS), electro-mechanical braking (EMB), and active suspension. This study focuses on the ball screw manufacturing workshop of Geely, a prominent enterprise in the new energy vehicle industry in China. The facility is equipped with advanced technologies, including intelligent machining centers, industrial robots, and automated guided vehicles (AGVs), with an annual output of approximately 48,000 pairs.

Despite its high level of automation, the workshop’s energy consumption and environmental impact had not been systematically assessed. Addressing this issue, the study first quantifies the energy costs and environmental impacts based on a proposed environmental cost model tailored for smart manufacturing systems. Subsequently, energy efficiency and carbon efficiency of manufacturing units are evaluated using the proposed environmental performance indicators. Finally, by integrating these indicators into an EC-VSM, the study visualizes the value stream, carbon emissions, and energy consumption patterns of the ball screw manufacturing system to support informed production decision-making.

As shown in Figure 8, the ball screw manufacturing system includes Sawing unit, Roughing turning unit, Roughing milling unit, Grinding unit 1, and Grinding unit 2, wherein Grinding unit 1 includes “Grinding center hole”, “Fine grinding external circle”, and Grinding unit 2 is “Fine grinding screw middle diameter”. They are the main production processes of the ball screw.

4.1. Basis Data

The data utilized in this study were collected from the ball screw smart manufacturing workshop of Geely, a new energy vehicle manufacturer in China. The data includes ball screw production indicators, equipment operational status across manufacturing units, the operating status of inter-unit material handling equipment, material consumption data, and information on factory facilities and intelligent service systems.

4.1.1. Data Acquisition and Pre-Processing

First, this study acquires time-series data on manufacturing systems’ environmental costs (e.g., equipment power, material, waste) via intelligent sensors and logistics systems. As shown in

Table 1. Data collection approaches.

Data	Data Sources	Measurement Devices	Sampling Frequency	Error Ranges
Machine power values	Technical specifications, direct/indirect measurement, theoretical/empirical models	Power sensors, data acquisition systems	Real-time sampling during the processing process, every day, monthly	$\pm 8\mathrm{\%}$
Carbon emission factors	Ministry of Ecology and Environment of the People’s Republic of China	https://www.mee.gov.cn	2024	$\pm 12\mathrm{\%}$
Electricity price	Enterprise Quarterly Electricity Bill Report	Internal company website search	2024, quarterly	$\pm 6\mathrm{\%}$
α in cutting-power model	Experimental calibration, theoretical/empirical models, algorithm derivation	Multi-channel power sensor, current transformer	Real-time sampling during the processing process	$\pm 5\mathrm{\%}$
Logistics transportation distance values	Logistics management system records, GPS track data	GPS positioning devices	Every time, weekly	$\pm 5\mathrm{\%}$
Load values	Weighing equipment records, logistics management system data	Electronic floor scales, goods identification system	Every time, weekly	$\pm 7\mathrm{\%}$

, it includes data sources, measurement equipment, sampling frequency, and error ranges for key variables such as equipment power, carbon emission factors, electricity prices, α in the cutting-power model, logistics transportation distance values, and load values.

Secondly, the acquired data are filtered to remove missing or invalid values, ensuring data reliability. Concurrently, the generalized extreme studentized deviate (GESD) test is employed to identify and detect outliers in the dataset, leveraging its adaptability across diverse application scenarios [63]. This approach enables the detection of outliers through a comparison between the studentized deviates $R$ of $n$ extreme observations and the critical value $\lambda$. Specifically, the extreme observations refer to those associated with the $n$ largest deviations from the mean value $x$. The $i$-th studentized deviate $R$ of extreme observation $x_i$ is calculated by Equation (30), the $i$-th critical value $\lambda$ is defined by Equation (31), and the tail area probability $p$ is defined by Equation (32).

$$R_i = \left|x_i-\overline{x}\right|/\sigma$$

(30)

$${\lambda }_i = {\displaystyle \frac{\left(n-i\right)t_{n-i-1,p}}{\sqrt{\left(n-i-1\right)\left(n-i-1 + t_{n-i-1,p}^2\right)}}}$$

(31)

$$p = \omega /\left[2\left(n-i-1\right)\right]$$

(32)

where $R_i$ is the $i$-th studentized deviate of $x_i$; $\sigma$ is the standard deviation. ${\lambda }_i$ is the $i$-th critical value. $t_{n-i-1,p}$ is the t-distribution with $n-i-1$ degrees of freedom. $p$ is the tail area probability. $\omega$ is the significance level. If $R_i$ is larger than ${\lambda }_i$, the $x_i$ is regarded as an outlier and is dropped from the dataset. Otherwise, the $x_i$ is not considered an outlier.

4.1.2. Basic Data Statistics

Table 2. Ball screw production indicator data.

Indicator	Monthly Demand (Piece)	Daily Production (Piece)	Cycle Time (h)
Data	4800	200	2

shows production parameters, including cycle times, monthly demand volumes, and daily output capacities for ball screws.

Table 3. Operating status data for processing equipment in each manufacturing unit.

Manufacturing Unit	Processor	Standby State	Idle State	Cutting State
		Average Standby Power $({\mathit{P}}_{\mathit{s}\mathit{t}}^{\mathit{p}}/\mathit{k}\mathit{W}$)	Average Idle Power $({\mathit{P}}_{\mathit{i}\mathit{d}\mathit{l}\mathit{e}}^{\mathit{p}}/\mathit{k}\mathit{W}$)	${\mathit{P}}_{\mathit{c}\mathit{u}\mathit{t}}/\mathit{k}\mathit{W}$
Sawing unit	Sawing machine	0.21 [0.20, 0.22]	0.4 [0.38, 0.43]	1.0 [0.93, 1.07]
Roughing turning unit	Turning machine	4.8 [4.56, 5.09]	11.2 [10.62, 11.94]	CNC machine cutting power: $P_{cut} = P_c + P_{ca}$ $\left\{\begin{array}{l}{P}_c\\ P_{ad}\end{array}\right\} = \left\{\begin{array}{l}{\displaystyle \frac{F_c\cdot V_c}{60}}\\ \alpha \cdot P_c\end{array}\right\}$ $\mathrm{where} F_C$$\mathrm{related} \mathrm{to} \mathrm{material} \mathrm{hardness}, \mathrm{tool} \mathrm{parameters}, \mathrm{and} \mathrm{depth} \mathrm{of} \mathrm{cut}/\mathrm{feed}; \alpha$$\mathrm{is} \mathrm{a} \mathrm{positive} \mathrm{coefficient}; \alpha = 0.1\sim 0.3$;
Roughing milling unit	Four-axis machining center	7.5 [7.13, 7.95]	19.8 [18.32, 21.12]
Grinding unit 1	CNC cylindrical grinding machine	6.0 [5.71, 6.36]	14.2 [13.46, 15.15]
	CNC grinding machine	4.3 [4.09, 4.56]	6.8 [6.44, 7.25]
Grinding unit 2	CNC thread grinding machine	5.2 [4.94, 5.51]	9.8 [9.19, 10.54]

details the power information of processing equipment in each manufacturing unit under different operational states.

Table 4. Operational status data for execution equipment in each manufacturing unit.

Manufacturing Unit	Executor	Standby State	Idle State	Transferring State
		Average Standby Power$({\mathit{P}}_{\mathit{s}\mathit{t}}^{\mathit{t}}/\mathit{k}\mathit{W}$)	Average Idle Power $({\mathit{P}}_{\mathit{i}\mathit{d}\mathit{l}\mathit{e}}^{\mathit{t}}/\mathit{k}\mathit{W}$)	Average Transfering Power $({\mathit{P}}_{\mathit{t}\mathit{r}}/\mathit{k}\mathit{W}$)
Sawing unit	ABB IRB 1300 robot	0.31 [0.30, 0.32]	1.3 [1.24, 1.36]	2.3 [2.15, 2.46]
Roughing turning unit	ABB IRB 1300 robot	0.31 [0.30, 0.32]	1.2 [1.14, 1.25]	2.3 [2.15, 2.46]
Roughing milling unit	ABB IRB 1300 robot	0.31 [0.30, 0.32]	1.2 [1.14, 1.25]	2.2 [2.05, 2.35]
Grinding unit 1	ABB IRB 1600 robot	0.30 [0.29, 0.31]	1.3 [1.24, 1.36]	2.0 [1.89, 2.14]
Grinding unit 2	ABB IRB 1600 robot	0.30 [0.29, 0.31]	1.2 [1.14, 1.25]	2.1 [1.96, 2.25]

details the power information of the execution equipment.

Table 5. Material handling equipment operating status data between manufacturing units.

Equipment	Standby State	Idle State	Load Transportation State
	Average Standby Power $\left({\mathit{P}}_{\mathit{s}\mathit{t}}\right)/\mathit{k}\mathit{W}$)	Average Idle Power $({\mathit{P}}_{\mathit{i}\mathit{d}\mathit{l}\mathit{e}}/\mathit{k}\mathit{W}$)	Average Load Transportation Power $({\mathbf{P}}_{\mathit{l}\mathit{o}\mathit{a}\mathit{d}}/\mathit{k}\mathit{W}$)
AGV	0.32 [0.31, 0.33]	1.1 [1.01, 1.20]	2.2 [2.07, 2.31]
Hoist	1.2 [1.12, 1.29]	2.0 [1.91, 2.12]	5.04 [4.92, 5.17]
Manipulator	1.0 [0.94, 1.06]	2.2 [2.10, 2.28]	4.2 [4.08, 4.33]

provides the power information of material handling equipment between manufacturing units.

Table 6. Operating power data for factory facilities and smart service equipment.

Equipment	$\mathbf{Average} \mathbf{Operation} \mathbf{Power} ({\mathit{P}}_{\mathit{A}\mathit{o}\mathit{p}}/\mathit{k}\mathit{W}$)	Variant Interval $({\mathit{P}}_{\mathit{A}\mathit{o}\mathit{p}}/\mathit{k}\mathit{W}$)
Exhaust system	8	[7.62, 8.46]
Lighting system	0.6	[0.57, 0.63]
Air compressor	11	[10.47, 11.62]
Intelligent terminal	0.3	[0.28, 0.32]

catalogs resource consumption metrics: raw material utilization, tool longevity, and coolant replenishment volumes per manufacturing unit.

Table 7. Material consumption data.

Information Items		Raw Material Consumption (kg)	Tool Life (min)	Cutting Fluid Replacement Volume (L)
Sawing	Saw-cutting	0.152	100	0.034
Roughing	Turning	4.154	48	0.224
	Milling	3.387	60	0.024
Grinding	Cylindrical grinding	2.575	80	0.025
	Grinding center hole	0.353	80	0.018
	Thread grinding	1.20	180	0.034

details the operational power of factory facilities and intelligent service equipment required to maintain the production environment. In addition, in accordance with the latest specification towards the enterprise greenhouse gas emissions accounting in China [64], the carbon emission factor for electricity ${CEF}_{elec}$ is $0.6205 \mathrm{k}\mathrm{g}{\mathrm{C}\mathrm{O}}_2\mathrm{e}/\mathrm{k}\mathrm{W}\mathrm{h}$. According to enterprise quarterly electricity bill report [4], electricity price $R_{elec}$ is 0.875 CNY/kWh.

4.2. Environmental Cost Calculation

4.2.1. Environmental Price Cost Calculation

The environmental price cost of the ball screw manufacturing system includes direct and indirect energy costs. According to the environmental cost model mentioned in Section 4, first, the environmental price cost of each manufacturing unit in the manufacturing system is calculated. Next, the environmental price cost of material handling between manufacturing units is calculated. Finally, the environmental price cost of factory facilities and intelligent service units that maintain the production environment in the workshop is calculated.

(1)

Manufacturing units

Roughing turning is a multi-equipment collaborative manufacturing unit, with the processor being a turning machine and the executor being an ABB IRB 1300 robot, and materials are transported by AGV. The environmental price cost calculation for roughing turning is as follows: (a) In roughing turning, there are two types of operation coupling: TEP and PET. The environmental price cost for the coupling stage is calculated using Formula 10. (b) The environmental price cost for the turning processing stage is calculated. The specific calculation process and results are shown in

Table 8. Environmental price cost calculation for roughing turning.

Stage	Calculate	Result
		Energy Consumption	Environmental Price Cost
Operational coupling stage (TEP, PEP, PET)	$EP{C}_{OC-SC} = R_{elec}\cdot {\displaystyle \sum _{i = 1}^n{\displaystyle \sum _{r = 1}^{m}SE{B}_r^i}}$ $(i =$ turning machine, four-axis machining center, ABB IRB 1300 robot, AGV)	0.254 kwh	CNY 0.21
Turning processing stage	$EP{C}_{M-Pro} = {\displaystyle \sum R_{elec}\cdot \left(\begin{array}{l}SE{B}_{standby}^{turning} + \\ SE{B}_{idle}^{turning} + \\ SE{B}_{cutting}^{turning}\end{array}\right)}$	1.870 kwh	CNY 1.64
total	$E_{\mathrm{total}} = EP{C}_{OC-SC} + EP{C}_{M-Pro}$	2.124 kwh	CNY 1.85

:

Similarly, the environmental price cost calculation results for each manufacturing unit are shown in

Table 9. Environmental price costs for each manufacturing unit.

Environmental Price Cost	Sawing	Turning	Milling	Grinding 1	Grinding 2	Total
${EPC}_{OC-SC}$	CNY 0.02	CNY 0.21	CNY 0.30	CNY 0.42	CNY 0.26	CNY 4.15
$EPC$	CNY 0.03	CNY 1.85	CNY 4.15	CNY 5.07	CNY 5.42	CNY 16.52

. The total environmental price cost for the ball screw manufacturing system unit is CNY 16.52.

(2)

Material handling between manufacturing units

In the ball screw manufacturing system, equipment used for material handling between manufacturing units includes AGVs, conveyor lines, etc. The environmental price cost of material handling was calculated according to the environmental cost model of transportation processes mentioned in Section 4.2, and the results are as follows. In the ball screw manufacturing system, the environmental price cost of material handling between manufacturing units is CNY 6.18.

$${EPC}_{tran} = R_{elec}·{EC}_{tran} = CNY6.18$$

(33)

(3)

Factory facilities and intelligent service units

The factory facilities that maintain the workshop production environment mainly include air compressors, lighting, and exhaust systems, while intelligent service equipment mainly includes office computers and smart terminals. According to the environmental cost model of auxiliary and intelligent services proposed in Section 4.3, the energy consumption of workshop factory facilities and intelligent service equipment was calculated, and the results are as follows. In the ball screw manufacturing system, the environmental price cost of factory facilities and intelligent service equipment that maintain the production environment is CNY 7.59.

$${EPC}_{Ind} = R_{elec}{\sum }_{b = 1}^{\xi }{\sum }_{i = 1}^n\left(P_{\xi }^{i,a} \times T_{\xi }^{i,a}\right) = CNY7.59$$

(34)

4.2.2. Environmental Impact Cost Calculation

Utilizing the environmental impact cost calculation model for manufacturing units proposed in Section 4.1, the energy–carbon emissions, material carbon emissions, and waste carbon emissions for each manufacturing unit are calculated. As shown in

Table 10. Carbon emissions for each manufacturing unit.

Environmental Impact Cost	Sawing Unit	Turning Unit	Milling Unit	Grinding Unit 1	Grinding Unit 2	Total
${EIC}_{elec}$	0.02 kg	1.32 kg	2.94 kg	3.60 kg	3.84 kg	11.72 kg
${EIC}_m$	0.02 kg	0.87 kg	1.49 kg	2.66 kg	2.3 kg	7.34 kg
${EIC}_w$	0.01 kg	0.56 kg	0.74 kg	1.29 kg	1.25 kg	3.85 kg
${EIC}_{total}$	0.05 kg	2.75 kg	5.17 kg	7.55 kg	7.39 kg	22.91 kg

, the carbon emissions information for each manufacturing unit is provided. The total carbon emissions for the manufacturing units of the ball screw manufacturing system amount to 22.91 kg, with energy–carbon emissions at 11.72 kg, material carbon emissions at 7.34 kg, and waste carbon emissions at 3.85 kg.

Section 4.2 and Section 4.3, which present the environmental cost model of transportation between manufacturing units and the environmental costs model of auxiliary and smart service units, calculate the environmental impact costs of material handling between units, factory facilities, and smart service equipment.

Table 11. Environmental cost for the ball screw smart manufacturing system.

Environmental Cost	Manufacturing Units	Transportation Between Manufacturing Units	Auxiliary and Smart Service Units	Total
$EPC$	CNY 16.53	CNY 6.18	CNY 7.59	CNY 30.3
$EIC$	22.91 kg	4.31 kg	5.40 kg	32.62 kg

shows the carbon emission information for the ball screw smart manufacturing system.

Three key sources of uncertainty compromise the accuracy of environmental cost accounting results: (i) data-level anomalies (e.g., instantaneous power peaks, sensor-acquired energy consumption fluctuations); (ii) contextual variability of core parameters (regional and seasonal changes in electricity carbon emission factor ${CEF}_{elec}$ and electricity price $R_{elec}$); and (iii) model structural uncertainties (e.g., linear simplification for multi-equipment coupled energy consumption).

To address these issues, this study adopts targeted uncertainty quantification approaches: probabilistic statistical methods convert fluctuations in contextual parameters (e.g., ${CEF}_{elec}$, $R_{elec}$) into environmental cost confidence intervals; error propagation theory quantifies systematic deviations in environmental costs caused by model simplifications; interval analysis assesses the impact of data anomalies and measurement errors on environmental cost estimates.

Quantitative results of parameter uncertainty are presented in

Table 12. Quantified results of parameter uncertainty.

Environmental Cost	Reference Value	95% Confidence Interval	Coefficient of Variation
${EPC}_{total}$	CNY 30.3	CNY (28.12, 32.56)	2.8%
${EPC}_{total}$	32.62 kg	(29.87, 35.42) kg	4.1%

.

4.2.3. Sensitivity Analysis of Environmental Costs in Ball Screw Manufacturing Systems

Electricity prices exert a direct impact on environmental costs, while electricity carbon emission factors serve as the core basis for the accurate accounting of indirect greenhouse gas emissions derived from electricity consumption. Both constitute critical input parameters in the environmental cost model. Owing to the regional heterogeneity in energy structures (e.g., differences in the proportion of fossil energy versus renewable energy) and electricity market mechanisms (e.g., market-oriented pricing or government-regulated pricing), significant variations in electricity prices and electricity–carbon emission factors exist across different countries and regions. Secondly, from the perspective of emission reduction economics, improving material utilization efficiency and reducing waste carbon emissions represent important pathways to achieve low-cost emission reduction.

Based on the results of environmental cost accounting for the ball screw manufacturing system, this study selected four core parameters—electricity price, electricity–carbon emission factor, material consumption, and waste carbon emission—to conduct a sensitivity analysis, aiming to reveal the extent to which fluctuations in each parameter affect environmental costs. The sensitivity coefficient of environmental costs to changes in different input parameters was calculated using Equation (35). This coefficient reflects the relative change in environmental costs caused by a unit relative change in the parameter, and the calculation formula is as follows:

$$S_i = {\displaystyle \frac{Y\left(X_i + {△X}_i\right)-Y\left(X_i\right)}{Y\left(X_i\right)}}$$

(35)

where $S_i$ is the sensitivity of the input variable $X_i$; $Y$ is output results; ${△X}_i$ is input variation value.

After varying the parameters, the results were compared with the initial values, as shown in Figure 9a. The electricity price had a relatively significant impact on environmental price costs, with a sensitivity coefficient of 0.050. As shown in Figure 9b, material consumption and waste carbon emissions exhibit minimal changes after fluctuation, indicating low sensitivity, whereas the carbon emission factor for electricity significantly impacts environmental impact cost, with a sensitivity coefficient of 0.026.

4.3. Environmental Cost Assessment

4.3.1. Environmental Cost Assessment of Ball Screw Manufacturing Processes

Building upon the environmental cost calculations for the ball screw manufacturing system (Section 4.2), the environmental cost of ball screw manufacturing is assessed according to the production-oriented environmental cost performance indicators mentioned in Section 3.3. The corresponding evaluation results are integrated into the ball screw manufacturing state description. The calculation of environmental cost assessment indicators is presented in Figure 10. The benchmark for these indicators is derived from the historical average values of the respective indicators, as detailed in

Table 13. Benchmarking guidance for manufacturing systems.

Indicators	Benchmark Value
${ER}_S^i$	56.2%
${\eta }_b$	$6.91 \mathrm{C}\mathrm{N}\mathrm{Y}/{\mathrm{C}\mathrm{O}}_2$
${ER}_M^i$	76.5%
${ER}_M^{OC}$	8.52%
${\eta }_p$	$2.95 \mathrm{p}\mathrm{i}\mathrm{e}\mathrm{c}\mathrm{e}/\mathrm{k}\mathrm{g}{\mathrm{C}\mathrm{O}}_2$

. These indicators offer valuable insights into the key areas that can be impacted through more effective management of energy consumption.

The EEe indicators are centered on energy efficiency, encompassing three specific metrics: Energy Utilization Rate ${ER}_S^i$, Value-Added Energy Consumption Ratio ${ER}_M^i$, and Operational Coupling Energy Consumption Ratio ${ER}_M^{OC}$. The calculation of these indicators depends on a set of variables, including equipment power values and the α parameter in the cutting-power model. Notably, the uncertainty inherent in these variables is propagated to the results of the indicators through the process of energy consumption calculation.

The CEe indicator, with a core focus on carbon efficiency, encompasses two specific metrics: economic return rate of carbon efficiency ${\eta }_b$, production rate carbon efficiency ${\eta }_p$. The calculation of these metrics hinges on a range of variables, such as carbon emission factors and logistics “equivalent values”. It is noteworthy that the uncertainty inherent in these variables exerts an influence on the indicator results through the process of carbon emission calculation.

As demonstrated in

Table 14. The sawing unit is employed as a case.

Manufacturing Units	Indicators	Reference Value	Key variables and Uncertainty Ranges	Uncertainty Propagation Results (95% Confidence Interval)
Sawing unit	${ER}_M^i$	57%	Machine power values (±8%), Logistics transportation distance values (±5%), Load values (±7%),	[54.2%, 59.8%]
	${\eta }_p$	$20 \mathrm{p}\mathrm{i}\mathrm{e}\mathrm{c}\mathrm{e}/\mathrm{k}\mathrm{g}\mathrm{C}\mathrm{O}$	carbon emission factors (±12%), Logistics transportation distance values (±5%), Load values (±7%)	$[18.5 \mathrm{p}\mathrm{i}\mathrm{e}\mathrm{c}\mathrm{e}/\mathrm{k}\mathrm{g}\mathrm{C}{\mathrm{O}}_{\mathrm{2}}$$, 21.7 \mathrm{p}\mathrm{i}\mathrm{e}\mathrm{c}\mathrm{e}/\mathrm{k}\mathrm{g}\mathrm{C}{\mathrm{O}}_{\mathrm{2}}$]

, the sawing unit is employed as a case example to elaborate on the propagation mechanism of such uncertainties between the EEe and CEe indicators.

Likewise, as presented in

Table 15. Reference values and results of uncertainty propagation.

Manufacturing Units	Indicators
	${\mathit{E}\mathit{R}}_{\mathit{M}}^{\mathit{i}}$	${\mathit{\eta }}_{\mathit{p}}$
Sawing unit	57% [54.2%, 59.8%]	$20 \mathrm{p}\mathrm{i}\mathrm{e}\mathrm{c}\mathrm{e}/\mathrm{k}\mathrm{g}\mathrm{C}{\mathrm{O}}_{\mathrm{2}}$$[18.5 \mathrm{p}\mathrm{i}\mathrm{e}\mathrm{c}\mathrm{e}/\mathrm{k}\mathrm{g}\mathrm{C}{\mathrm{O}}_{\mathrm{2}}$$, 21.7 \mathrm{p}\mathrm{i}\mathrm{e}\mathrm{c}\mathrm{e}/\mathrm{k}\mathrm{g}\mathrm{C}{\mathrm{O}}_{\mathrm{2}}]$
Turning unit	76% [72.1%, 79.8%]	$3.64 \mathrm{p}\mathrm{i}\mathrm{e}\mathrm{c}\mathrm{e}/\mathrm{k}\mathrm{g}\mathrm{C}{\mathrm{O}}_{\mathrm{2}}$$[3.37 \mathrm{p}\mathrm{i}\mathrm{e}\mathrm{c}\mathrm{e}/\mathrm{k}\mathrm{g}\mathrm{C}{\mathrm{O}}_{\mathrm{2}}$$, 3.88 \mathrm{p}\mathrm{i}\mathrm{e}\mathrm{c}\mathrm{e}/\mathrm{k}\mathrm{g}\mathrm{C}{\mathrm{O}}_{\mathrm{2}}$]
Milling unit	75% [71.3%, 78.6%]	$2.33 \mathrm{p}\mathrm{i}\mathrm{e}\mathrm{c}\mathrm{e}/\mathrm{k}\mathrm{g}\mathrm{C}{\mathrm{O}}_{\mathrm{2}}$$[2.15 \mathrm{p}\mathrm{i}\mathrm{e}\mathrm{c}\mathrm{e}/\mathrm{k}\mathrm{g}\mathrm{C}{\mathrm{O}}_{\mathrm{2}},$$2.52 \mathrm{p}\mathrm{i}\mathrm{e}\mathrm{c}\mathrm{e}/\mathrm{k}\mathrm{g}\mathrm{C}{\mathrm{O}}_{\mathrm{2}}$]
Grinding unit 1	77% [73.5%, 80.4%]	$2.93 \mathrm{p}\mathrm{i}\mathrm{e}\mathrm{c}\mathrm{e}/\mathrm{k}\mathrm{g}\mathrm{C}{\mathrm{O}}_{\mathrm{2}}$$[2.69 \mathrm{p}\mathrm{i}\mathrm{e}\mathrm{c}\mathrm{e}/\mathrm{k}\mathrm{g}\mathrm{C}{\mathrm{O}}_{\mathrm{2}}$$, 3.15\mathrm{p}\mathrm{i}\mathrm{e}\mathrm{c}\mathrm{e}/\mathrm{k}\mathrm{g}\mathrm{C}{\mathrm{O}}_{\mathrm{2}}$]
Grinding unit 2	79% [75.6%, 82.3%]	$3.52 \mathrm{p}\mathrm{i}\mathrm{e}\mathrm{c}\mathrm{e}/\mathrm{k}\mathrm{g}\mathrm{C}{\mathrm{O}}_{\mathrm{2}}$$[3.28\mathrm{p}\mathrm{i}\mathrm{e}\mathrm{c}\mathrm{e}/\mathrm{k}\mathrm{g}\mathrm{C}{\mathrm{O}}_{\mathrm{2}},$$3.77 \mathrm{p}\mathrm{i}\mathrm{e}\mathrm{c}\mathrm{e}/\mathrm{k}\mathrm{g}\mathrm{C}{\mathrm{O}}_{\mathrm{2}}$]

, the reference values and results of uncertainty propagation (with a 95% confidence interval) of the manufacturing unit are detailed as follows.

4.3.2. Improvements

Given that the ${ER}_S^i$ of the ball screw manufacturing system is only 55.7%, which is lower than the benchmark value of 56.2%, and the sawing unit, turning unit, and milling unit all fall below the benchmark value of 76.5%, these indicators suggest significant optimization potential for manufacturing systems.

Energy–carbon efficiency of manufacturing unit assessment identifies the sawing and roughing milling units as critical bottlenecks. System-level assessment reveals excessive auxiliary and material handling energy consumption, indicating significant wastage.

The following improvements were made over one month based on the above assessment results. The manufacturing units were merged, the manufacturing process was optimized, and four manufacturing units were established: cutting, roughing, grinding, and thread grinding. The product transfer warehouse and parts warehouse were optimized, and the inspection process was merged. Finally, the logistics layout and scheduling of material handling equipment were optimized to shorten the logistics distance between manufacturing units and reduce the idling rate of material handling equipment.

4.3.3. Future-State Description

As shown in Figure 11, the original five manufacturing units were consolidated into four manufacturing units. The environmental costs of the consolidated sawing unit, roughing unit, and manufacturing system were calculated using the proposed environmental cost calculation model and reevaluated based on the proposed production-oriented environmental cost performance indicators.

4.4. Results

The enhanced ball screw manufacturing system demonstrates notable improvements across multiple performance dimensions. Comparative results between before and after improvement are presented below:

(1)

At the operational level

(i)

As shown in

Table 16. Operational coupling energy consumption ratio of each manufacturing unit.

${\mathit{E}\mathit{R}}_{\mathit{M}}^{\mathit{O}\mathit{C}}$	Sawing	Turning	Milling	Grinding 1	Grinding 2
	66.67%	11.95%	7.18%	8.32%	4.74%

, the sawing unit reached an astonishing 66.67%, far exceeding the benchmark value of 8.52%. After implementing improvement measures for one month, the ${ER}_M^{OC}$ of the manufacturing unit decreased from 66.67% to 53%. However, there still exists a significant gap compared to the benchmark value.

(ii)

Roughing unit: Compared to the ${ER}_M^{OC}$ of the rough turning unit (11.95%), it exceeded the benchmark value of 8.52%. But with the rough milling unit (7.18%), the combined rough machining unit achieved 7.90%. It fell below the benchmark value of 8.52%, following the manufacturing units being merged.

(2)

At the manufacturing unit level

(i)

Sawing unit: The energy efficiency of the manufacturing unit increased from 57% to 63%.

(ii)

Roughing unit: Compared to the energy efficiency of the rough turning unit (76%) and the rough milling unit (75%), the combined rough machining unit achieved 78%, representing a certain improvement in unit energy efficiency.

(3)

At the manufacturing system level

(i)

Reduction in logistics equivalent: From 1002.4 $\mathrm{k}\mathrm{g}·\mathrm{m}$ to 627.2 $\mathrm{k}\mathrm{g}·\mathrm{m}$, a decrease of 37.4% is shown in Table 17. This indicates that through optimized logistics layout and process management, the efficiency of material handling and transportation has been significantly improved. The reduction in the logistics equivalent directly correlates with lower logistics costs and also reduces production delays caused by logistics bottlenecks.

(ii)

Reduction in total energy consumption: Total energy consumption decreased from 34.62 kWh to 30.46 kWh, a reduction of 12.0%.

(iii)

Reduction in total carbon emissions: Total carbon emissions decreased from 32.62 kg to 30.09 kg, a reduction of 7.8%.

(iv)

Improved system energy efficiency: System energy efficiency improved from to, an increase of 8.6%. This indicates that the proportion of energy used to directly create product value in the total energy consumption of the manufacturing system has increased.

(v)

Improved system return rate carbon efficiency: System return rate carbon efficiency improved from 55.7% to 58%. This indicates that the economic benefits of the manufacturing system have reduced carbon emissions on the environment and improved the performance of the manufacturing system.

Table 17. Comparison and analysis of data before and after improvement.

Indicator	Before Improvement	After Improvement	Benefits
Logistics equivalent	1002.4 kg.m	627.2 kg.m	37.4%
Total energy consumption	34.62 kWh	30.46 kWh	12.0%
Total carbon emissions	32.62 kg	30.09 kg	7.8%
Manufacturing system energy efficiency	55.7%	58%	4.1%
Manufacturing system return rate carbon efficiency	6.89 $\mathrm{C}\mathrm{N}\mathrm{Y}/{\mathrm{k}\mathrm{g}\mathrm{C}\mathrm{O}}_2$	7.48 $\mathrm{C}\mathrm{N}\mathrm{Y}/{\mathrm{k}\mathrm{g}\mathrm{C}\mathrm{O}}_2$	8.6%

Table 17. Comparison and analysis of data before and after improvement.

Indicator	Before Improvement	After Improvement	Benefits
Logistics equivalent	1002.4 kg.m	627.2 kg.m	37.4%
Total energy consumption	34.62 kWh	30.46 kWh	12.0%
Total carbon emissions	32.62 kg	30.09 kg	7.8%
Manufacturing system energy efficiency	55.7%	58%	4.1%
Manufacturing system return rate carbon efficiency	6.89 $\mathrm{C}\mathrm{N}\mathrm{Y}/{\mathrm{k}\mathrm{g}\mathrm{C}\mathrm{O}}_2$	7.48 $\mathrm{C}\mathrm{N}\mathrm{Y}/{\mathrm{k}\mathrm{g}\mathrm{C}\mathrm{O}}_2$	8.6%

5. Discussion

5.1. Comparison of the EC-VSM Method with Related Work

By conducting a comparative analysis from three dimensions, namely objective, process, and result, we aim to demonstrate its superiority over other related works. As shown in Table 18, EC-VSM has the following advantages over other related works:

(1)

Objective feasibility: EC-VSM takes the integration of carbon efficiency as its core objective. At the logistics level, it explicitly incorporates the accounting of energy consumption and carbon emissions between manufacturing units, while both E2 VSM and EVSM lack analysis in this aspect. At the production–logistics integration level, it incorporates the operational coupling effects of multiple devices into the carbon efficiency analysis framework, effectively addressing the deficiency of traditional methods in depicting the collaborative mechanism between the two, thus being more in line with the actual characteristics of complex manufacturing systems.

(2)

Comprehensiveness of process: The analysis process of EC-VSM demonstrates stronger systematicness. In terms of operational coupling modeling, it innovatively proposes the MECMfg-OC model to characterize the coupling relationships among multiple devices, which is significantly superior to the non-coupling analysis of production-oriented E2 VSM and the focus on single-machine energy characteristics of energy-oriented EVSM. In the integration of environmental costs, it constructs a structured model including EPC and EIC, overcoming the ambiguity in cost division of production-oriented E2 VSM and the one-sidedness of analysis of energy-oriented EVSM. In visualization technology, it realizes the collaborative display of value flow, carbon flow, and energy flow, and integrates dual indicators of energy efficiency and carbon efficiency, which is more suitable for the needs of comprehensive decision-making compared with the limited-dimensional display of E2 VSM and EVSM.

(3)

Guidability of results: The practical verification of EC-VSM shows significant advantages. In case applications, it achieves a 12.0% reduction in total energy consumption, a 7.8% reduction in carbon emissions, and an increase in system energy efficiency to 58%, forming a multi-dimensional improvement verification. At the level of improvement guidance, it provides specific paths for unit consolidation and logistics optimization through the identification of bottleneck units, which is more practical than the broad suggestions of other methods and has more prominent decision support value.

Table 18. Comparing the EC-VSM method with other approaches against.

Comparison Dimensions		Economic and Environmental Value Stream Maps (E2 VSM) [65]	Energy Value Stream Maps (EVSM) [9]	EC-VSM (This Paper)
Objective	Production	Analyze dynamic material, energy, and information flows in multi-product manufacturing systems to assess economic and environmental performance.	Identify main energy consumers in production lines and evaluate energy efficiency.	Integrate carbon efficiency into production processes.
	Logistics	Does not account for logistics energy consumption between units.	Neglect logistics energy consumption and carbon emissions between manufacturing units.	Explicitly include logistics energy consumption and carbon emissions between manufacturing units.
	Production–logistics integration	Ignore multi-device operational coupling effects in production–logistics integration.	The carbon efficiency analysis does not address the coupling between production and logistics.	Integrates multi-device operational coupling effects into carbon efficiency analysis.
Process	Operational coupling modeling	Do not consider modeling operational coupling between multiple equipment.	Focus on individual machine energy profiles without modeling operational coupling between devices.	Propose MECMfg-OC model to characterize multi-equipment operational coupling using state EnergyBlocks and component EnergyBlocks.
	Environmental cost integration	Integrate direct/indirect energy consumption and related emissions but lack a structured division of environmental costs.	Visualize energy flows and identify energy waste in production processes.	Integrate EPC and EIC into the environmental cost model.
	Visualization	Display value-added/non-value-added time, energy, and emissions for each product.	Visualize energy flows and identify energy waste in production processes.	Simultaneously display value flow, carbon flow, and energy flow, integrating energy efficiency and carbon efficiency metrics.
Result	Case validation	Achieved simulation error <10% in a brake pad manufacturing case; identified high non-value-added (NVA) energy consumption.	Quantified energy reduction opportunities (42% and 50%) in medical device and pharmaceutical factories.	Total energy consumption decreased by 12.0%, carbon emissions reduced by 7.8%, and system energy efficiency improved from 55.7% to 58%.
	Improvement guidance	Enabled what-if analysis (e.g., comparing gas vs. electric ovens) to optimize energy use.	Guided operational changes (e.g., adjusting chiller controls) to reduce auxiliary energy consumption.	Identified bottleneck units via EC-VSM, guiding unit consolidation and logistics optimization.

Table 18. Comparing the EC-VSM method with other approaches against.

Comparison Dimensions		Economic and Environmental Value Stream Maps (E2 VSM) [65]	Energy Value Stream Maps (EVSM) [9]	EC-VSM (This Paper)
Objective	Production	Analyze dynamic material, energy, and information flows in multi-product manufacturing systems to assess economic and environmental performance.	Identify main energy consumers in production lines and evaluate energy efficiency.	Integrate carbon efficiency into production processes.
	Logistics	Does not account for logistics energy consumption between units.	Neglect logistics energy consumption and carbon emissions between manufacturing units.	Explicitly include logistics energy consumption and carbon emissions between manufacturing units.
	Production–logistics integration	Ignore multi-device operational coupling effects in production–logistics integration.	The carbon efficiency analysis does not address the coupling between production and logistics.	Integrates multi-device operational coupling effects into carbon efficiency analysis.
Process	Operational coupling modeling	Do not consider modeling operational coupling between multiple equipment.	Focus on individual machine energy profiles without modeling operational coupling between devices.	Propose MECMfg-OC model to characterize multi-equipment operational coupling using state EnergyBlocks and component EnergyBlocks.
	Environmental cost integration	Integrate direct/indirect energy consumption and related emissions but lack a structured division of environmental costs.	Visualize energy flows and identify energy waste in production processes.	Integrate EPC and EIC into the environmental cost model.
	Visualization	Display value-added/non-value-added time, energy, and emissions for each product.	Visualize energy flows and identify energy waste in production processes.	Simultaneously display value flow, carbon flow, and energy flow, integrating energy efficiency and carbon efficiency metrics.
Result	Case validation	Achieved simulation error <10% in a brake pad manufacturing case; identified high non-value-added (NVA) energy consumption.	Quantified energy reduction opportunities (42% and 50%) in medical device and pharmaceutical factories.	Total energy consumption decreased by 12.0%, carbon emissions reduced by 7.8%, and system energy efficiency improved from 55.7% to 58%.
	Improvement guidance	Enabled what-if analysis (e.g., comparing gas vs. electric ovens) to optimize energy use.	Guided operational changes (e.g., adjusting chiller controls) to reduce auxiliary energy consumption.	Identified bottleneck units via EC-VSM, guiding unit consolidation and logistics optimization.

5.2. Results Discussion

(1)

This study proposes models for operational coupling phenomena and energy consumption calculation in multi-equipment collaborative manufacturing. These models reveal the distribution patterns of operational coupling environmental costs. Case analysis shows the proposed EC-VSM identifies inefficient links arising from multi-equipment coupling and overcomes the limitations of the E2VSM, which focuses on a single product, and the EVSM, which is confined to a single machine. For example, the sawing unit reached an astonishing 66.67%, far exceeding the benchmark value of 8.52%. The result indicates that inefficient coordination is a major source of environmental costs in manufacturing systems. It also validates the necessity of the proposed coupling energy consumption model.

(2)

At the manufacturing unit level, merging rough machining units increased system energy efficiency. Efficiency rose from turning (76%) and milling (75%) to 78%, more than the benchmark value of 76.5%. The core mechanism lies in eliminating redundant processes caused by material transfer between equipment. Examples include idle equipment and waiting times. This increases the proportion of direct collaborative work between processors (machine tools) and actuators (robots). These findings validate that reducing non-value-added coupling links can improve system energy efficiency.

(3)

At the system level, the proposed environmental cost evaluation metric effectively identifies inefficiencies in manufacturing processes. It provides a basis for driving energy–carbon synergy optimization in smart manufacturing systems. This metric integrates environmental and operational dimensions. It offers a new perspective for jointly addressing the environmental challenges and operational efficiency needs of smart manufacturing systems. Existing research often focuses on greenhouse gas emissions [66] while neglecting their underlying energy consumption [67]. This model simultaneously quantifies both greenhouse gas emissions and their corresponding energy consumption.

The proposed methodology focuses on assessing environmental impact and energy consumption in smart manufacturing units. It helps factories develop targeted environmental improvement strategies. It also evaluates system-level environmental impacts. This supports the sustainable development of smart manufacturing systems.

6. Conclusions

It is of vital significance for achieving sustainable smart manufacturing that the assessment of energy consumption and carbon efficiency of the stage of multi-equipment collaborative manufacturing operational coupling is conducted.

First, to address the specific stage of multi-equipment collaborative manufacturing operational coupling in the production process, as well as the multi-state characteristics of equipment operation, multidependencies among operational states, multi-source nature of carbon emissions, and spatiotemporal sequence coupling, this study proposes a production execution logic model for MECMfg-SC, which effectively explains the multi-equipment operational coupling mechanism. Subsequently, the concepts of “state energy consumption unit” and “component energy consumption unit” are introduced to characterize the energy consumption features under different equipment states, thereby revealing the energy consumption characteristics in the operational coupling stage and realizing accurate calculation of energy consumption during this stage. Based on these modules, an environmental cost accounting model for smart manufacturing systems is constructed, and verification in a ball screw manufacturing unit demonstrates that the proposed model exhibits efficient and reliable performance, providing a valuable reference for energy management and environmental cost control in multi-equipment collaborative manufacturing scenarios.

Second, an environmental cost model is established for smart manufacturing systems, covering manufacturing units, inter-unit material handling, factory infrastructure, and intelligent service equipment. This model enables comprehensive evaluation of both greenhouse gas emissions and associated energy consumption, providing a holistic framework for environmental cost assessment in complex manufacturing environments.

Finally, energy and carbon efficiency indicators for MECMfg-SC are integrated into the EC-VSM, which realizes the visualization of value streams, carbon flows, and energy consumption patterns. This integration facilitates the lean identification and management of energy–carbon hotspots, offering actionable guidance for producers, managers, and supervisors in formulating and implementing energy-saving and emission-reduction strategies.

The current validation is limited to the ball screw manufacturing system. Future research will extend this methodology to broader manufacturing scenarios. Focus will be placed on developing hybrid data–knowledge-driven environmental cost evaluation methods and intelligent unit-level assessment frameworks. Further work will also investigate system-level optimization strategies for carbon efficiency.