Eco-design improvement for the diaphragm forming process

Abstract

This paper proposes an eco-design method to suggest improvements over an existing diaphragm forming process (DFP). In the proposed method, a systematic procedure is developed to provide eco-design guidance to engineers, and it includes four steps. In Step 1, design functions are analysed through functional diagrams to provide the information of sub-functions and functional flows. In Step 2, the eco-design requirements are captured via quality functional deployment (QFD) and then translated to design functions via functional analysis (FA). Then, the design functions are prioritized according to the eco-design requirements in considerations. In Step 3, the prioritized design functions are used to generate possible design concepts through the morphological chart. In Step 4, the generated concepts are assessed based on the energy use and process time of DFP. The utility of the proposed method is to adapt quality tools for continual improvements in the context of eco-design. An existing DFP is used for this work to demonstrate and validate the method’s applicability. The result of this research shows that the integration of QFD and FA can systematically guide the generation of new eco-design concepts for DFP with less dependence of design intuitions.

Keywords:

• Eco-design
• quality functional deployment
• conceptual design
• functional analysis
• morphological chart
• diaphragm forming process

1. Introduction

Sustainable manufacturing (or environmentally conscious manufacturing) has become an important issue in a modern society (Duflou et al. 2012; Haapala et al. 2013). To reduce environmental impacts, this paper focuses on the application of eco-design techniques for manufacturing applications. Specifically, eco-design is a design strategy with a specific focus on environmental concerns, and it pays attention to the life cycle assessment (LCA) of a product (Abele, Anderl, and Birkhofer 2005; Fiksel 2009).

In general, three methodical phases can be classified in eco-design development: concept generation, environmental assessment and decision-making. Owning to the influence of LCA, environmental assessment has been the most studied, and the common metrics include energy consumption and material usage (Ashby 2013). Besides LCA, quality function deployment (QFD) is another common assessment tool that prioritizes engineering characteristics based on customer requirements.

Based on the assessment results, decision-making is applied to determine the course of actions. In a general design context, engineers may need to choose one design concept out of several options for further product development. At the point, they need to perform certain trade-off analysis to make the decision (e.g. energy consumption versus cost). For example, Jiang et al. (2014) have developed a decision plan method for remanufacturing process selection. Also, Kim et al. (2014) have presented sustainability improvement in manufacturing process based on a decision-guidance framework.

Comparatively, concept generation tends to be less explored in the eco-design research since it involves less tangible yet creative elements in the methodology development. Some examples include the guideline-based approach in Design for Environment (Fiksel 2009) and TRIZ (Theory of Inventive Problem Solving) adopted for eco-design (Russo, Rizzi, and Montelisciani 2014; Sakao 2007). Notably, adopting some design methodology for eco-design is not often easy. For example, the users of TRIZ require specialized training to apply the technique (Russo, Rizzi, and Montelisciani 2014).

In the domain of design theory and methodology, the ‘form-follows-function’ principle is quite well recognized but it has not been fully explored for eco-design. In design literature, ‘function’ is defined as a transformation that highlights the design purpose in view of specific inputs and outputs (Otto and Wood 2001) To emphasize the transformation, it is common to describe a function using a format: verb + noun. For example, the function of an electric motor can be described as ‘convert electrical energy to mechanical torque’, where electrical energy is the functional input and mechanical torque is the output. Alternately, the function of a motor can be briefly described as ‘convert energy’. By focusing on the motor’s function, the design engineer can ‘unlock’ their mind from the object ‘motor’ and start to brainstorm the possible solutions for achieving the function ‘convert energy’.

In the context of eco-design concept generation, the ‘form-follows-function’ principle has two advantages. Firstly, its basic concept is not difficult, and domain experts can learn and apply it with less challenge. Secondly, this principle has supported other concept generation strategies such as morphological chart (Pahl and Beitz 2007) and concept combination table (Ulrich and Eppinger 2012). These works have demonstrated the utility of ‘function’ to explore new design ideas without locking the brainstorming process on the ‘form’ (or object).

In the methodical development, it is intended to keep QFD for mapping eco-design requirements in the design process. The research inquiry is how to extend the framework of QFD to include design functions for concept generation. In this work, functional analysis (FA) is used to extract functional information of existing design (Otto and Wood 2001). Accordingly, one key aspect of the proposed method is to integrate QFD and FA to support eco-design concept generation. Specifically, through the QFD mapping, the proposed method will prioritize the design functions so that the design efforts can be sensibly deployed on eco-design improvement.

To promote the method’s applicability, this paper chooses the diaphragm forming process (DFP) as the eco-design application. One reason is that this process has a reasonable level of technical complexity, and new conceptual ideas for eco-design improvement are not entirely obvious. Then, the utility of the proposed method can be illustrated by generating eco-design concepts through a systematic procedure (rather than intuitions). In sum, two specific research objectives of this paper are set and described below.

•	Objective 1: Develop a concept generation method by integrating QFD and FA.
•	Objective 2: Propose eco-design concepts for the DFP.

The rest of this paper is organized as follows. Section 2 provides literature review. Section 3 provides an overview of the DFP. Section 4 provides the eco-design improvement methodology. Section 5 shows the application of the methodology to the DFP. Section 6 concludes the paper.

2. Literature review

To achieve product quality in view of customer satisfaction, one essence of QFD is to explicitly relate the customer’s concerns to engineering metrics (or characteristics) (Akao 1990). In this way, engineers are required to consider how the improvement of engineering metrics can directly lead to better customer satisfaction. In QFD, this mapping from customer requirements to engineering metrics is done via the relational matrix in the house of quality.

In the context of sustainable engineering, QFD has been extended by considering ‘environment’ as one type of customers. The relevant works include Green QFD (Cristofari, Deshmukh, and Wang 1996; Dong, Zhang, and Wang 2003; Zhang 1999), Environmentally Conscious QFD (ECQFD) (Kaebernick, Kara, and Sun 2003) and QFD for Environment (QFDE) (Devanathan et al. 2010; Masui et al. 2003; Sakao 2007). In addition, QFD has been applied for facilitating remanufacturing. Yüksel (2010) has used QFD to incorporate the information from refurbishing facilities for the remanufacturing of automobile engines. Jiang et al. (2014) have used fuzzy linear regression to support the selection of remanufacturing process plans in the QFD framework.

In industrial applications, Vinodh and Rathod (2010) have applied QFDE for the design of rotary switch, and Rathod, Vinodh, and Madhyasta (2011) have integrated ECQFD and LCA for the design of electric vehicles. In these two papers, the generation of eco-design concepts is based on QFD I and II to identify the parts that are essential for improving eco-performance. Instead of focusing on parts and components, this paper utilizes FA (Otto and Wood 2001) and the morphological chart (Pahl and Beitz 2007) to support eco-design concept generation. In brief, FA is concerned with the technical transformation of a design so that engineers can focus on the ‘purpose’ (rather than the ‘form’) of the design to simulate new design ideas (i.e. the ‘form-follows-function’ principle). The morphological chart is a table form that allows designers to list possible solutions during the brainstorming stage for individual functions.

Towards the effort of sustainable manufacturing in this paper, it is recognized that certain amounts of research have been conducted, as reviewed by Duflou et al. (2012) and Haapala et al. (2013). One research direction is to conduct LCA for manufacturing processes. For example, Overcash, Twomey, and Isaacs (2009) have calculated the energy and mass losses using engineering guidelines and industrial experience. Also, a LCA framework has been proposed for analysing the inventory of manufacturing unit process (Gibovic and de Ciurana 2008; Kellens et al. 2012). Related to manufacturing processes, one general concern is to reduce their energy consumptions. Haapala et al. (2012) have developed a process modelling method to improve the environmental performance of metal casting. It is claimed that the decrease of the electricity consumption can be obtained via preheating the material scrap, which is required for charging the furnace (Jacobus et al. 2001). As reviewed in these works, beyond the efforts of modelling and evaluations, the generation of eco-design ideas for manufacturing processes usually relies on the observations by domain experts. In this context, this paper is intended to apply the concept generation technique from design methodology in the manufacturing domain.

This paper is concerned with the DFP. In literature, researchers have extended the practice of LCA to the forming processes of the sheet metal (Ingarao, Di Lorenzo, and Micari 2011). Their concerns are to use materials and energy more efficiently. The investigated processes were based on experimental works and concluded that using different materials can lead to various material and energy savings. Gutowski, Dahmus, and Thiriez (2006) reported that the energy requirements for the manufacturing process are not the same, due to the different use of process rate. They have also suggested an approach to reduce the energy use by redesigning the manufacturing process. Also, Paralikas, Salonitis, and Chryssolouris (2013) proposed the energy efficiency indicator method to estimate the energy consumption of the cold roll forming process. In contrast, the proposed method of this paper is intended to improve the eco-performance of DFP by formally generating new ideas for the manufacturing process. This concept generation effort for DFP has not been reported in literature as observed by the authors.

3. The diaphragm forming process

3.1. Overview

Diaphragm forming is a method used for the production of aerospace parts. Initially, the method is applied on a thermoplastic matrix composite. Then, the method is extended to the application to the thermosetting matrix composite (Bersee et al. 2007). In this study, the material, which is considered in the DFP, is as general as a polymer matrix composite. DFP is conducted based on thermal energy, which is normally obtained from electrical energy. Thus, the key physical component is the heating source. The detailed information of the components of DFP is explained in the next subsections.

The methods by which the forming process is accomplished can be classified into basic types such as compression moulding, vacuum forming and diaphragm forming (Long 2007). In general, DFP has the advantages compared to other forming processes such as the quality of the produced components. Yet, it also has drawbacks such as long cycle time (Tucker 1997), which may lead to environmental impacts. This issue motivates the application of eco-design improvement for the original DFP in this study.

DFP consists of two main phases: the forming process and the curing process. In the forming process, a composite prepreg (or laminate) is first placed between two sheets (i.e. diaphragms) which are made from silicon rubber. The sheets are well clamped to avoid any defects on the formed shape. Then, the heater is turned up to warm up the material until the temperature distribution is guaranteed to be distributed evenly on the material. Next, the hydrostatic pressure is applied on the material to create the deformation of the diaphragms (Smiley and Pipes 1988). When the air between two diaphragms is removed by a pump, the diaphragm is forced to form the polymer to the required shape. In addition, the diaphragms are stretched into the cavity of the mould under the atmosphere pressure. Finally, the part (i.e. the material) can be safely received after removing the two diaphragms. The part is then required to be placed in an autoclave for the purpose of curing (treatment). Figure 1 illustrates the main steps of the DFP.

Figure 1. Schematic of diaphragm forming process.

In the curing process, the part is placed under a vacuum and then pressurized during the heated cure cycle. The high pressure on the part helps to minimize resin (polymer) voids and achieve the desired temperature. Autoclaves normally operate from 120 to 230 °C at 7 bars of pressure. Most common materials cured in an autoclave are advanced composites such as carbon fibre and epoxy resins. The curing cycle time can be estimated according to the used material and is generally between 1.5 and 12 h.

3.2. Eco-design aspect

To perform eco-design improvement, one aspect is to analyse the major components of DFP and try to find better design solutions for the existing DFP. In this study, the components of the existing DFP are listed as follows.

•	Heating element: used as a heating source to soften the material for the final shape.
•	Diaphragm: during the forming process it translates the geometric deformation to the prepreg/laminate and controls the forming stage. It is made of silicon rubber and is able to survive the high processing temperatures (e.g. from 350–400 °C) without rupture.
•	Vacuum pump 1: used to draw the air between the two diaphragms. The maximum vacuum pressure is about 100 KPa.
•	Vacuum pump 2: used to draw the air from the vacuum box. The maximum vacuum pressure is about 100 KPa.
•	Vacuum box: the part containing the mould and contains the air that needs to be drawn out.
•	Mould: the part that uses to form the shape.
•	Film: the material that is made from nylon or reusable silicon rubber. It is used to help the removal of the material from the mould.
•	Autoclave: used to finish the complete polymerization. The autoclave temperature is above the polymer melting temperature and the autoclave pressure is up to 7 bar.

Given the existing DFP, it is observed that DFP is a relatively long process, wherein the consumption of the energy is high. Human input is essential and required for most of the stages of the process, and hence the time of the process has become long. Those observed problems incur impacts on the environment. In this case, the eco-design requirements are to reduce the cycle time and make the process more environmentally benign.

In the next section, an eco-design improvement method will be proposed and applied to the DFP. To better control the scope of the work, only the first step (i.e. the forming process) is included in this study. The design question here in practice is how to improve the existing forming process (as illustrated in Figure 1) for reducing environmental impacts. The next section is intended to address this question via a systematic method.

4. Methodology

The proposed method for eco-design improvement has four main steps, which are illustrated in Figure 2. In this methodical workflow, the existing design (or manufacturing process) is first analysed via FA (Step 1) and quality planning (Step 2). FA yields the mapping of design functions and components, while quality planning based on QFD yields the requirements, metrics and components of the existing design. Then, concept generation (Step 3) takes the functional and QFD-related information to generate multiple design concepts using the morphological chart. The last step is to evaluate the eco-performance of proposed design concepts. The new idea of the proposed method is to map the eco-design requirements to the design functions, which are then used for generating new design concepts. These methodical steps are further explained in this section.

Figure 2. Workflow of the proposed method for eco-design improvement.

4.1. Functional analysis

As discussed earlier, the key idea of a function is to focus on the transformation between inputs and outputs. The role of FA is to analyse design functions systematically. In this aspect, three types of flows for functional inputs/outputs are specified: material, energy and information (Otto and Wood 2001). Figure 3 shows these three types of flows for the DFP. The basic function of DFP is to convert a prepreg to a formed part, representing the material flow in bolded arrows in Figure 3. It is expected that this DFP requires energy, thus the energy flow in narrow arrows. The information flow, represented in dashed arrows, signals the precedence and status of the manufacturing process.

Figure 3. Three types of flows for the diaphragm forming process.

The box in Figure 3 can be viewed as a ‘black box’, in which engineers can design more specific details for the DFP. Otto and Wood (2001) have suggested the identification of design functions via process descriptions. Let F = {f₁, f₂, … } be a set of design functions. By following the manufacturing steps described in Section 3.1, six functions (symbolized as f₁–f₆) for the DFP are listed as follows.

•	f1: import prepreg
•	f2: tighten diaphragm
•	f3: soften prepreg
•	f4: shape prepreg
•	f5: harden part
•	f6: export part

Notably, these functions are described using the format ‘verb + noun’. Also, these functions have the material/energy/information flows as functional inputs/outputs. As an example, Figure 4 illustrates the function, f₃, that takes the flow of heat energy to ‘soften the prepreg’. The information flow indicates the signals of heat on/heat off. This functional description leaves some space for engineers to brainstorm the possible heating methods as an innovation process for eco-design improvement. By applying a similar analysis, the six functions of the DFP can be further developed and connected via their input/output flows, resulting in a functional diagram.

Figure 4. Illustration of functional description.

To understand the existing design better, a function-component matrix is also constructed to show which existing components are used to achieve the functions. Let FC = [fc_ij] be such a matrix, and C = {c₁, c₂, … } be the set of design components. If the component c_j is required to achieve the function f_i, then the matrix entry, fc_ij, is equal to 1 (otherwise, fc_ij = 0). Further information of the functional diagram and FC for the DFP will be provided in Section 5.

4.2. Quality planning

Beyond customer satisfaction, the consideration of environmental issues has been another goal in the pursuit of quality for companies (McCarty, Jordan, and Probst 2011). One common approach is to modify some existing quality tools for the eco-design purpose, and QFD has been commonly applied in this context. The initial effort can be found in the development of Green QFD (G-QFD) (Cristofari and Wang 1996). Subsequently, some methodical extensions have been presented such as extending G-QFD to QFD II by integrating QFD and LCA to evaluate different design concepts (Rathod, Vinodh, and Madhyasta 2011; Zhang 1999).

In our view, the essence of QFD is the mapping between customer requirements and engineering metrics (Akao 1990). While customer requirements highlight what customers want, the QFD mapping can associate which engineering metrics are important in view of customers. In this context, let R = {r₁, r₂, … } and M = {m₁, m₂, … } be the sets of customer requirements and engineering metrics, respectively. Then, let RM = [rm_ij] be the mapping matrix that corresponds to QFD. Usually, a 9-point scheme is applied to quantify rm_ij in QFD. For example, if r_i is strongly associated with m_j, rm_ij is equal to 9. For a medium or weak association, rm_ij is equal to 3 and 1, respectively. This scheme of QFD is applied in this paper.

In literature, RM is referred to the first mapping in QFD (or QFD I). Further mapping has also been proposed as QFD II, which maps from engineering metrics to design components (Akao 1990). Recall that C = {c₁, c₂, … } is the set of design components. Let MC = [mc_ij] be the mapping matrix between engineering metrics and design components. The same 9-point scheme is also applied to express the strength of association between the metric m_i and the component c_j as the common practice in QFD I and QFD II.

In sum, the intent of quality planning is to develop the mapping matrices RM and MC by following the practice of QFD. In the context of eco-design improvement, the environment is interpreted as one type of customers, and its requirements can be expressed as a subset of R, which can be further mapped to M and C via RM and MC, respectively. The specifics of the DFP will be presented in Section 5.

4.3. Concept generation

Concept generation here is carried out in two phases. The first phase is to utilize the matrix information (i.e. FC, RM and MC matrices) to prioritize the design functions identified in Section 4.1. In the second phase, the morphological chart is applied to the functions of high priority to generate design concepts for eco-design improvement.

The key technique of the first phase is matrix multiplication. As a logical chain, if the requirement r_i is strongly associated with the metric m_j (expressed in RM) and m_j strongly associated with c_k (expressed in MC), then r_i should be strongly associated with the component c_k. This logical chain can be formally determined via matrix multiplication (Hamraz, Caldwell, and Clarkson 2013). Given that ‘eco-requirements’ are provided, the issue is to determine which functions are strongly associated with these eco-requirements for eco-design improvement. Let RF = [rf_ij] be the mapping matrix between requirements and functions. Then, this matrix can be determined via the formulation below (where the superscript T means matrix transpose).(1)

To explain the use of RF, suppose that one requirement, r₁, expresses ‘less energy use’. Through the determination of RF, it is found that r₁ has the strongest association with f₃ ‘soften prepreg’. With this information at hand, if the engineers want to improve the existing product in view of r₁ ‘less energy use’, they should devote the efforts to think about alternate ways for achieving f₃ ‘soften prepreg’. By prioritizing the functions, engineers can rank their efforts in what aspects of the product that should be explored for new solutions.

Given the prioritized functions, the morphological chart is applied for concept generation. The key of this chart is to ask the engineers the possible solutions for achieving each function. In this way, engineers can focus on one function at a time, thus promoting the possibility of new ideas. To synthesize a design concept, engineers can pick one solution for each function based on their expertise and judgements. Notably, the morphological chart is well recognized as one concept generation method, and interested readers may refer to Lo, Tseng, and Chu (2010) for more details of the morphological chart.

4.4. Concept evaluation

Multiple design concepts can be obtained from the morphological chart. At this point, several concept selection methods from literature can be applied such as multi-criteria decision-making (Dieter and Schmidt 2009). In this paper, instead of determining a single design ‘winner’, we focus on two eco-design aspects of the DFP: energy use and processing time. Then, new design concepts are evaluated via the estimation of these two performance metrics. Further details can be provided in Section 5.

5. Methodical application to the diaphragm forming process

In this section, we will apply the methodology in Section 4 to the DFP and propose design concepts for eco-design improvement.

5.1. Functional analysis and quality planning

In FA, the DFP is first analysed according to the manufacturing procedure described in Section 3, and these procedural steps are listed on the first column of Table 1. The actual function(s) of each step are then developed using graphical representation, shown in the second column of Table 1. Based on the existing DFP, the components that are used to perform the functions are also included in Table 1.

Table 1. Functional analysis of DFP.

Procedure of DFP	Graphical representation of functions	Components
Put the prepreg between two diaphragms. Clamp the diaphragms.		c1: diaphragm
		c2: clamp
Remove air between diaphragms		c3: pump 1
		c4: tube
Heat the prepreg		c1: diaphragm
		c5: steel-coil heater
Remove air in the mold		c1: diaphragm
		c4: tube
		c6: pump 2
		c7: mold
		c8: vacuum box
Cool the mold and remove the formed part		c1: diaphragm
		c2: clamp
		c7: mold
		c8: vacuum box

To construct the overall functional diagram, the individual functions in Table 1 are connected according to their inputs and outputs, and the result is provided in Figure 5. In this functional diagram, the labels of information flows are omitted for simplicity, while the labels of material and energy flows are the same as those in Table 1. Then, this functional diagram can be viewed as an ‘exploded’ view of the functional black box in Figure 3.

Figure 5. Overall functional diagram of DFP.

In addition, based on Table 1, the FC matrix of the DFP can be constructed, and it is provided in Equation (2(2) ). To illustrate, we can check f₁, which is the first row of the FC matrix. This row shows that f₁ is related to c₁ and c₂, which corresponds to the information in Table 1 (i.e. diaphragm (c₁) and clamp (c₂) are used for ‘f₁: import prepreg’).(2)

In quality planning, it is intended to construct two matrices: RM and MC. In the context of eco-design, suppose that the engineers want to reduce energy use and processing time in the DFP. Accordingly, two eco-design requirements (r₁ and r₂) and metrics (m₁ and m₂) are proposed below, along with the RM matrix in Equation (3(3) ).

•	r1: less energy use
•	r2: less processing time
•	m1: energy consumption (in kWh)
•	m2: processing time (in minutes)

(3)

In this RM, we have rm₁₁ = rm₂₂ = 9 for the direct measure of energy and processing time. Then we have rm₁₂ and rm₂₁ = 3 since longer processing time implies more energy use (and vice versa). Notably, we do not list all requirements and metrics that are involved in a typical DFP. As we try to focus on eco-design, only the relevant requirements and metrics are listed for highlighting the product ‘hotspot’ for eco-design improvement.

Afterwards, MC is developed based on the analysis of the existing DFP, and this MC matrix is provided in Equation (4(4) ). The matrix entries with the value of ‘9’ indicate that the steel-coil heater has the strongest influence towards the energy use and processing time. Since the pumps (c₃ and c₆) also consume energy (not as much as a heater though), we have mc₁₃ = mc₁₆ = 3. Also, since the mould (c₇) can affect the dissipation of heat, it can impact the processing time (thus, mc₂₇ = 3). Other matrix entries with the value of ‘1’ indicate that the relevant components can slightly affect the energy use and processing time.(4)

Thus far, the matrices in Equations (2(2) )–(4(4) ) represent the mapping information from the existing DFP. Specifically, QFD yields the mappings of ‘requirements → metrics → components’, while FA yields the mapping of ‘functions → components’. To apply the ‘form-follows-function’ principle in the next subsection, the mapping of ‘requirements → functions’ will be derived via Equation (1(1) ), and it will be used to support the generation of eco-design concepts for DFP.

5.2. Concept generation and evaluation

After the matrices RM, MC and FC are obtained, we can determine the matrix RF using Equation (1(1) ) to prioritize the functions for eco-design improvement. The resulting RF matrix is provided in Equation (5(5) ). From this RF, it is found that the function f₃ (i.e. soften prepreg) has the strongest influence for achieving r₁ and r₂ (i.e. rf₁₃ = 117 and rf₂₃ = 111). Comparatively, f₁ and f₆ (i.e. import prepreg and export part) are not so important for r₁ and r₂, and thus, they are not included in the morphological chart for concept generation.(5)

Accordingly, the morphological chart is built in Figure 6 with the descending priority of functions based on the resulting RF. Since f₃ has the strongest influence, this function is primarily considered for generating new design ideas. First of all, while heating is unavoidable to soften the prepreg, it is noted that the ceramic-coil heater can potentially have better heating efficiency. Also, in view of the heating strategy, it may be efficient to adopt some ‘preheat’ strategy for the subsequent softening process. As one result, it is proposed to preheat the mould using an electric heater. Without introducing a new heater, it is noted that f₃ has an output ‘dissipated heat’ (as well as f₅), and it is proposed to use the dissipated heat to preheat prepreg 2 (i.e. the next prepreg prepared for the forming process).

Figure 6. Morphological chart.

Other design ideas are also generated by checking other functions (i.e. f₄ and f₂). Since these functions have less impact on r₁ and r₂, we can choose to devote less effort for brainstorming new ideas for these functions. Here, the essence of the proposed eco-design method is to prioritize the design functions so that the engineers can choose to devote their design efforts accordingly.

After developing the morphological chart, two design concepts are proposed accordingly. First of all, as the ceramic-coil heater and the single pump with rewired tube are quite feasible in practice, these two solution concepts are used for the two concepts. To differentiate, the first concept (namely, concept x) adopts the electric heater concept to preheat the mould. Alternately, the second concept (namely, concept y) adopts the dissipated heat to preheat prepreg 2 using a new fan. Both concepts x and y are illustrated in Figures 7 and 8, respectively.

Figure 7. Design concept x.

Figure 8. Design concept y.

To evaluate design concepts, we use two metrics defined earlier: energy use (m₁) and processing time (m₂). To estimate the energy use, we first find the power ratings of the components that consume energy and then approximate their usage times. To estimate the processing time, we check the time required for achieving the functions from f₂ (i.e. tighten prepreg) to f₅ (i.e. harden part). The functions of f₁ and f₆ are not considered here since the new concepts should have about the same processing time with the original design in these two functions.

Table 2 tabulates the initial analysis. In the column of energy use, the components that consume energy are listed for the original and proposed designs. In the column of processing time, the times of the original design are treated as a benchmark, and the times of concept x and y are commented with the note ‘same’ or ‘shorter’.

Table 2. Initial analysis of energy use and processing time.

	Energy use	Processing time
Original	Consumed by sheet-coil heater and two pumps	Tighten time → soften time → shape time → harden time
Concept x	Consumed by ceramic heater, electric heater, and one pump	Tighten time (same) → soften time (shorter) → shape time (same) → harden time (same)
Concept y	Consumed by ceramic heat, fan, and one pump	Tighten time (same) → soften time (shorter) → shape time (same) → harden time (shorter)

To perform the estimation, the power ratings and run times of the components in the DFP are recorded in Table 3. This table has two parts. The first part records the power ratings and run times of existing components, while the second part provides the estimations related to the proposed components in concepts x and y. Then, the energy use for concept evaluation can be determined by multiplying power rating with run time.

Table 3. Information for estimating energy use and processing time.

Components that consume power	Power rating (kW)	Run time (min)	Comments
Steel-coil heater	6.0	15.0	Used in original (soften prepreg)
Pump 1	0.56	7.8	Used in original (tighten prepreg)
Pump 2	0.74	9.6	Used in original and concepts x & y for shape prepreg
Pump 2 (Concepts x & y)	0.74	7.8	Concepts x & y: tighten prepreg
Ceramic-coil heater (Concepts x)	5.7	12.0	Used in concept x, shorter time due to preheat mold
Ceramic-coil heater (Concepts x)	5.7	14.0	Used in concept y, shorter time due to preheat prepreg 2
Electric heater (Concepts x)	2.0	5.0	Used in concept x, preheat mold during tighten prepreg
Fan (Concept x)	0.05	21	Used in concept y, cool mold during shape prepreg

Table 4 records the energy use and processing time of the original design and concepts x & y. First of all, the energy use is addressed by working on the function f₃ via the ceramic-coil heater (for less power rating) and the preheat strategy (for less heating time). By using only one pump (that has higher power rating), concepts x and y require more energy for f₂ (tighten diaphgram). Regarding the processing time, the fan is expected to reduce about 30% of the hardening time, which becomes 21 min in concept y. The trade-off, of course, is the extra energy use for the fan. Since fans generally have much lower power ratings (as compared to heating components), they represent a good option for engineers to reduce the processing time.

Table 4. Comparison of metrics of new design concepts.

	Original		Concept x		Concept y
	Energy (kWh)	Time (min)	Energy (kWh)	Time (min)	Energy (kWh)	Time (min)
f2: tighten diaphragm	0.073	7.8	0.096	7.8	0.096	7.8
f3: soften prepreg	1.5	15.0	1.3	12.0	1.4	14.0
f4: shape prepreg	0.12	9.6	0.12	9.6	0.12	9.6
f5: harden part	0	30.0	0	30.0	0.018	21.0
Total	1.693	62.4	1.516	59.4	1.634	52.4

5.3. Discussion

By reflecting on the new design concepts x and y, this section is intended to examine how the proposed method can actually support the generation of new concepts for eco-design improvement. By checking Table 4, it is observed that both concepts x and y can potentially reduce the energy use and processing time of the existing DFP. This result is mainly based on three design techniques, which are listed and commented below.

•	Use a more efficient heater for ‘soften prepreg’. This design technique is rather ‘obvious’, and the energy-saving result is quite predictable.
•	Adopt a ‘preheat’ strategy. This design technique is less obvious as preheat is not one of the original functions. Yet, by realizing that a fixed amount of energy is more or less required to the function ‘soften prepreg’, we tend to explore the energy consumption beyond the original heater, and thus the preheat strategy emerges in the process of using a morphological chart (i.e. Figure 6).
•	Adopt a fan. This design technique emerges when we consider the possible solutions for the function ‘harden part’ (see the morphological chart in Figure 6). This idea does not look interesting preliminarily until the dissipated heat is observed in the functional diagram in Figure 5. Combining the idea of preheat, the fan can actually redistribute the dissipated heat for some useful purpose.

As discussed here, we obtained the ideas of ‘preheat’ and ‘fan’ based on the reasoning process within the proposed methodology. Arguably, some talented engineers may come up with some similar eco-design ideas without using the proposed methodology. Yet, in reality, we cannot rely on the ‘light-bulb’ moments from few talented engineers entirely. Therefore, the proposed methodology is intended to support engineers to organize design information (in terms of matrices) and direct their design efforts systematically. In this way, it is expected that new eco-design ideas can emerge steadily as one kind of continual improvement emphasized in quality (or six-sigma) management (McCarty, Jordan, and Probst 2011).

6. Closing remarks

Sustainability in manufacturing is the issue of this study, in which a new methodology is proposed to support eco-design improvement at the early design stage. In the methodical approach, QFD is integrated with FA to prioritize the design functions for concept generation. An existing DFP is used in this study to demonstrate the proposed methodology. Accordingly, two concepts for DFP are generated based on the design functions to provide some design insights and evaluate the concepts in view of energy use and processing time. The practicality of this research lies in the context that DFP is an ordinary forming process. Then, the research approach of this paper can be referenced and applied to generate new eco-design ideas for other forming and manufacturing processes.

The contributions of this work towards advancing sustainable design and manufacturing practices are twofold. Firstly, as the ‘form-follows-function’ principle is well recognized in the domain of design methodology, it is applied in the QFD framework so that practitioners in sustainable design can follow and apply the proposed method. Secondly, the demonstration of DFP can encourage practitioners to apply the proposed method (especially the process of FA) for analysing and improving the eco-performance of manufacturing processes. In future work, experimental works for the forming process will be implemented for assessing the environmental issues including the curing stage. Besides, specific composite polymers such as carbon fibre and fibre-reinforced thermoplastics will be considered in the analysis of the forming process to identify their environmental impacts.

Disclosure statement

No potential conflict of interest was reported by the authors.

Funding

This work was supported by Libyan-North American Scholarship Program.

Acknowledgement

We would like to express our sincere thanks and appreciation to Mr. Hassan Alshahrani, a PhD student at Concordia University, Montreal, Canada, for providing us data and information of the DFP.