Reducing design uncertainty through model-based collaborative design methods when introducing new technologies: A Solomon four-groups empirical design study

doi:10.21203/rs.3.rs-2101538/v1

Download PDF

Research Article

Reducing design uncertainty through model-based collaborative design methods when introducing new technologies: A Solomon four-groups empirical design study

https://doi.org/10.21203/rs.3.rs-2101538/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

When introducing new technologies in product design, failing to identify and address uncertainties regarding technology feasibility, and the way technology introduction impacts the whole product development process, hinder the establishment of appropriate design requirements during conceptual stages. Ill-defined design requirements often lead to unpleasant and expensive surprises that arise late on the product development process, rendering the introduction of a new technology an expensive and time-consuming endeavor.

Uncertainties about the new technology can be known (information designers know is missing), or unknown (information designers do not know is missing).

Previous research by the authors has proposed a digital design platform and respective design methods to support cooperative efforts to identify and model uncertainties during conceptual phases. In this article, the usefulness of such platform is evaluated through an empirical study using a Solomon Four-Groups design study featuring the design of satellite components for additive manufacturing. The results of the study suggest that the proposed digital platform and associated design methods are useful for identifying uncertainties and for proposing measures to address them, through a cooperative modeling environment.

Digital design platform

design uncertainties

manufacturing engineering

testability

empirical study

Solomon four-groups design study

One of the most powerful strategies to ensure market permanence is the introduction of new technologies (Porter, 1985; Passemard and Kleiner, 2000).

However, new technology introduction is often hampered by uncertainties regarding technology capabilities, and the way the new technology impacts the whole product development process (Stroud, 2002). Failing to assess new technology capabilities and its impact on the overall product development process (PDP) activities, can lead to the establishment of ill-defined product design specifications and, consequently, a final product that fails to deliver the promised business benefits (Stroud, 2002; Goldberg et al., 2018).

The consequences of ill-defined design specifications can range from expensive and time-consuming redesign loops to technology abandonment (Thompson et al., 2016).

To avoid these consequences, there is a need to increase the knowledge about the new technology and reduce the number of uncertainties during conceptual phases (Dordlofva and Törlind, 2020).

In general, uncertainties can be described as known uncertainties (information designers know is missing, known unknowns, KU), or unknown uncertainties (information designers do not know is missing, unknown unknowns, UU) (Jensen et al., 2017). Strategies for dealing with known unknowns include test campaigns and assessments by experts (Sutclife and Sawyer 2013; Ramasesh and Browning, 2014; Dordlofva and Törlind, 2020). However, unknown unknowns are difficult to find as practitioners do not know what information is missing (Jensen et al., 2017), as a consequence, UU are the source of unpleasant surprises during late PDP where design changes are the most expensive. Failing to account for UU can lead to the establishment of non-adequate project budget for contingencies (Raydugin, 2012).

Previous research (REF) by the authors have proposed model-based design methods for uncertainty identification and modelling based on multidisciplinary, holistic product representations compiled in a digital design platform, and a technique of design constraints modelling and replacement.

The digital platform has been developed as a boundary object where different product representations (such as function models and activity models) can interact and support design endeavors and knowledge transfer among practitioners with different expertise and perspectives.

The proposed methods and design platform have been applied to specific case studies before, yielding positive results in the identification of uncertainties, both known and unknown; however, it is yet to be determined if the results obtained are due to the methods application or to the influence of confounding variables (Sawilowsky, 1994; Trochim, 2021).

To assess the usefulness of the proposed methods and digital design platform for uncertainty identification and modelling, a necessary step for their validation according to authors such as Pedersen et al. (2000) or Frey et al. (2006), an experimental set up inspired on the Solomon four-group design study was implemented to qualitatively prove or refute the following hypotheses:

Hypothesis 1: The proposed multidisciplinary, holistic design methods have a positive effect on the identification, during conceptual phases, of design uncertainties (both known and unknown) related to the feasibility of a new technology and its impact on later product development activities.

Hypothesis 2: The modelling environment supports the transformation of the identified uncertainties into known knowns.

2.1. Uncertainties during technology introduction

Introducing new technologies in their products, companies aim at market permanence (Porter, 1985; Passemard and Kleiner, 2000).

However, not every technological change is strategically convenient, technology increases competitive advantage if it has a significant role in reducing overall costs and increasing product uniqueness or performance (Porter, 1985).

The successful assessment of a new technology introduction and the eventual establishment of design requirements for its implementation is often hampered by uncertainties (Mavris et al., 1999; Goldberg et al., 2018). These uncertainties are related to technology capabilities (what the technology is and is not capable of, such as what can or cannot be manufactured when introducing a new manufacturing technology), and the way the technology introduction impacts the whole product development process regarding its activities, resources and schedule (Porter, 1985; Stroud, 2002; Browning and Ramasesh, 2015).

The solution to this problem is to reduce the number of uncertainties, increasing the knowledge about the new technology and its impact on the PDP (Dordlofva and Törlind, 2020). Literature (Browning and Ramasesh, 2015) divides uncertainties in two categories, known uncertainties, information designers know is missing, known unknowns (KU) and unknown uncertainties, information designers do not know is missing, unknown unknowns (UU).

Strategies for dealing with known unknowns are well-studied and implemented on literature and on industrial contexts, they often include prototyping (Elverum and Welo, 2015; Jensen et al., 2017), test campaigns, or assessments by experts (Sutclife and Sawyer 2013; Ramasesh and Browning, 2014; Dordlofva and Törlind, 2020). However, unknown unknowns are difficult to find as practitioners do not know what formation is missing (Browning and Ramasesh, 2015; Jensen et al., 2017). Moreover, UU are sometimes facts that practitioners know but did not come to their minds when designing (Ramasesh and Browning, 2014).

If the UU are left unknown, the unknown data gap about the new technologies might be filled with carried over knowledge from previous design projects, due to the tendency to design products with similar features to their predecessors (Kumke et al., 2016; Seepersad et al., 2017). As previous data is not always applicable to new design contexts, ill-defined design specifications can be established, leading to the development of products that do not fulfill their full potential (Kumke et al., 2016; Seepersad et al., 2017). Other, more serious consequences are the establishment of non-adequate project budget for contingencies (Raydugin, 2012) and the development of products that fail to deliver business benefits (Mavris et al., 1999; Goldberg et al., 2018).

Sutcliffe and Sawyer (2013) presented a review of existent techniques to deal with uncertainties. Their work indicates that techniques such as interviews, observations, workshops, scenarios, or prototypes are quite mature and well-established. These techniques can, however, be expensive or time consuming specially if they are not systematic and well-planned.

Prototyping strategies, as those proposed by Jensen et al. (2013), Borgue et al. (2019), or Dordlofva and Törlind (2020) are adequate for transforming KU into KK, however, they seldomly result on the discovery of UU. In the work by Borgue et al. (2019), for example, the authors use prototypes for shedding light into manufacturing uncertainties for additive manufacturing technologies (AM) in the context of a flow connector for satellite applications. During this process, they uncovered an UU, related to the difficulty of removing the flow connector from the building platform; for this reason, an appropriate new design specification was established to address this difficulty. Nevertheless, the discovery of this UU was accidental and relying on prototyping for UU discovery is not time or cost effective.

The work by Ramasesh and Browning (2014) and Browning and Ramasesh (2015) presents a list of factors thar foster UU, and several strategies for UU identification. On essence, techniques for UU identification call for a purposeful and systematic UU identification procedure based on modelling strategies, communication, and an alertness culture. These statements are aligned with those from Sutcliffe and Sawyer (2013), who also stated the importance of models and problem decomposition for UU identification. However, as these authors point out, there is still room for improvements in what concerns modelling strategies for UU identification.

Multidisciplinary product development modelling strategies and performance simulations are widely used for uncertainty assessment and reduction (Mavrik, 1999; Struck and Hensen, 2007; Ogaji et al., 2007; Goldberg et al., 2018) during conceptual stages. These strategies are tailored for KU reduction and for mitigating their effect on the overall PDP, however, they are in general not concerned with the identification of UU, especially those related to design specifications. Some approaches from the software industry, such as the one proposed by Henricksen and Indulska (2004), focus on UU identification through user-based object modelling approaches to model information quality and manually flag UU. Their graphical representation is well suited for developer and user based UU identification and specification of requirements of an application. These approaches, however, are not applied to the architecture design of hardware.

In general, there is a lack of systematic model-based (Sutcliffe and Sawyer, 2013; Browning and Ramasesh, 2015; Dordlofva and Törlind, 2020) platforms to systematically identify, model and solve design specifications uncertainties.

2.2. A multidisciplinary design platform to model uncertainty and identify unknown unknowns

Previous work by the authors (REF) have proposed the implementation of a multidisciplinary digital design platform for identifying, modelling, and solving design specifications uncertainties through the collaborative efforts of practitioners with different expertise.

This digital design platform is based on product architecture representations realized on enhanced function models (FM). This platform is intended to act as a boundary or intermediary object (Boujut and Blanco, 2003) to foster collaboration among practitioners with different expertise and support design exploration and decision making.

The function models the platform is based on, present a hierarchical decomposition of a product architecture from the main product functional requirement (FR) to the lowest level FRs to ensure product performance (Claesson, 2006). One design solution (DS) is assigned to each FR, representing the function´s physical embodiment in the product architecture (Fig. 1). Moreover, constraints are linked to the DSs, to reduce design space and specify the boundaries inside which the DS should be realized. In the case of new technology introduction, and in this model representation, constraints are related to technology capabilities and testing in the context of the PDP.

To evaluate the impact that different product architectures have on the PDP, the platform links its functional representation with PERT diagrams that model the different PDP activities. This activity modelling strategy is particularly interesting for modelling test activities, as test and design phases are intertwined throughout the PDP (Tahera et al., 2019).

Moreover, to enable schedule assessment and optimization, the risk of not performing a certain activity (or performing it partially) is included on PERT diagram and on the portion of the FM related to that activity.

In Fig. 1, the FM in which the product architecture model is based on is presented along the respective PERT diagram. An example featuring a flow connector for satellite applications and its respective validation activities is included in the Fig. 1 as well, to illustrate the connection of the FM with the activity models and risk assessment.

To identify UU related with new technology introduction, a process of constraints replacement is proposed. As design processes rarely start from scratch, many design and development processes with new technologies start from an old product that implements a previous, well-established technology. In the process of constraints replacement, the FM of the product with the previous technology is reused, and the constraints related to the old technology are removed and replaced with their counterparts for the new technology. In the example from Fig. 1 the flow connector is traditionally machined from a metal block, and the constraints in the function model are related to this manufacturing technology. To redesign the flow connector for, for example, additive manufacturing, the machining constraints are removed and replaced with their AM counterparts, as presented in Fig. 2.

Through this constraints replacement technique practitioners are encouraged to systematically analyze how a new technology would affect a product architecture and PDP activities. This way, new technology KU can be included in the architecture analysis during conceptual phases. Moreover, the constraints replacement procedure enables the discovery of UU related to technology feasibility and the effect the technology has on the PDP.

The constraints replacement technique and the design platform have been applied to specific case studies before, yielding positive results in the identification of design specifications uncertainties; however, it is yet to be determined if the results obtained are due to the methods application or to the influence of confounding variables (Sawilowsky, 1994; Trochim, 2021).

2.3. Experimental design methods for evaluating the usefulness of a method

One of the simplest methods for evaluating the usefulness of a method or a tool is a two-group posttest-experiment (Fig. 3.a) (Trochim, 2021). In this arrangement, one group applies the method or tool to perform a task, and the other group does not. The results are them compared based on preestablished metrics. However, this method does not provide a way to compare both groups’ baselines, so the results from these studies could be due to intrinsic differences between the two groups. For instance, when evaluating the usefulness of a design tool for reducing conceptual design times, the two groups could represent two different companies. If one company uses the tool and the other does not, it is not possible to know if the differences obtained in design times are due to the tool or the companies.

One way to establish a baseline to compare both groups is performing a pretest-posttest experiment (Fig. 3.b) (Trochim, 2021). In this arrangement, both groups perform a pretest, to establish their baseline results to be then compared with the results obtained after applying the method or tool of interest. When evaluating the design times of two companies, a pretest would establish a comparison of the current (without the tool) design times of each company.

Nevertheless, a drawback of this configuration is the difficulty to establish if the obtained results are due to the method, or the “practice” obtained during the pretest.

The Solomon Four-Group Design (Fig. 3.c) (Sawilowsky, 1994; Trochim, 2021) is an experiment arrangement designed to deal with the effects of the pretest. This arrangement requires four groups, two of the groups use the method or tool and two do not. Furthermore, two of the groups receive a pretest and two, do not.

This experiment design is more complex to set up and analyze due to the number of groups involved, however, it reduces the internal validity issues, enables the researcher to exert better control over the experiment variables, and to make sure that the pretest did not influence the experiment results (Sawilowsky, 1994; Trochim, 2021; May et al., 2020).

To evaluate the previously proposed holistic design platform and the technique of constraint replacement, the Solomon-four group design experiment was implemented in the context of a Swedish research project with industrial partners.

This project aimed at demonstrating the development of a digital infrastructure and platform for the industrialization of Additive Manufacturing (AM). Industrial case representatives, digital environment providers and research institutes were brought together as representative stakeholders to the introduction of AM in industrial contexts. The objective of the project was to evaluate current solutions enabling digital data traceability and digital modelling and simulation activities for optimal design of product and process definition.

In this research project, the proposed design platform served two objectives: (1) Evaluate the platform´s usefulness to identify UU and propose measures to cope with them; (2) Showcase design capabilities developed through four years of previous research.

To evaluate the platform´s usefulness, the Solomon-four group design experiment was preferred, as it has been proved advantageous to reduce the effect that confounding variables and external factors have on the study results (Trochim, 2021).

A purposeful, homogeneous sampling (Palinkas et al., 2015) with two associative hypotheses in a qualitative study context (Chigbu, 2019), was preferred.

A total of 12 industrial practitioners participated in the study in six groups of two participants. Moreover, a test run was performed at the beginning of the study with two extra participants to adjust experiment details and improve instructions. Further details of the experiment and its participants are found in the next section.

The Solomon-four group design experiment showcased the design platform to industrial practitioners as well. Moreover, after the study was finished, its results were presented to the research project participants, which enabled a further discussion, included in the “Results” section of this article.

3.1. Solomon-four group design experiment

The experiment was designed to probe or refute two hypotheses:

Hypothesis 1

The proposed multidisciplinary, holistic design methods have a positive effect on the identification, during conceptual phases, of manufacturability and testability unknown unknowns related to the implementation of a new technology.

Hypothesis 2

The proposed multidisciplinary, holistic design methods have a positive effect on the establishment of measures to address the identified uncertainties.

In this experiment, two design case studies are implemented where practitioners evaluate the redesign for AM of two products which are part of a fluid management system of a satellite: (1) a flow connector and, (2) a heat exchanger.

To limit the scope of the study, the function models of the analyzed products contained only constraints related to manufacturability and testability. Moreover, the PERT diagrams used in the study have just included the system´s validation activities, which in the context of space components are comparable with the qualification or acceptance phases. Validation activities are those performed to control if the product fulfills the intended function and business case.

The Solomon-four group design experiment was performed twice (Iteration 1 and Iteration 2), each time with three groups of two participants each. On the first iteration, the flow connector was used the pre-test, and the heat exchanger as the test. On the second iteration, the roles were inverted, and the heat exchanger was used as the pre-test and the flow connector as the test. This way, the need for Group D was eliminated, as its results are obtained from the Group A of the opposite iteration (Fig. 4). Moreover, interchanging the pre-test and test increases results reliability, as it eliminates the concerns of the methods results being due to the test of choice.

Table 1

Experiment participants´ company roles and experience (ys)*
	Group	Participant	Current company role	Total industrial experience	Total experience working with AM	Experience in professional roles as a designer	DfAM professional experience	Experience in professional roles involving test planning or test performance
Iteration 1	A	1A1	Business analyst	4	1	0	0	0
	A	1A2	CEO	5	8 months	6 months	6 months	3 months
	B	1B1	AM Research engineer	13	3.5	6 months	1.5	0
	B	1B2	Lead product developer	6	6	-	-	6
	C	1C1	Engineer lead for AM	3	2	0	2	0
	C	1C2	Software Engineer	4	4	0	0	0
Iteration 2	A	2A1	R&D engineer	+ 20	2.5	+ 20	2.5	+ 15
	A	2A2	Researcher in data science	8	3	0	0	0
	B	2B1	R&D engineer	27	3	0	3	0
	B	2B2	Technical project leader	15	6	3	3	3
	C	2C1	Researcher	3.5	3.5	0	3	7
	C	2C2	Technical product responsible	24	3	24	3	0

3.2. Case studies for pre-tests and tests

The two products implemented in the experiments are part of a fluid management system of a satellite. A branch of a fluid management system is made of seven components, schematized in Fig. 5: (1) Heat exchanger, (2) Pressure sensor, (3) Temperature sensor, (4–6) Pressure vessel A, B and C, (7) Flow connector.

The flow connector and the heat exchanger were redesigned for AM, the experiment participants were presented with the task of analyzing those redesigns (weather during a pre-test or a test) and voice their concerns related to component manufacturability and testability. The concerns that practitioners mention are interpreted as uncertainties.

3.3. Activities performed during the experiments

The experiment was performed twice, iteration 1 (I1) and iteration 2 (I2), with six participants in each iteration, divided in three groups, Group A, Group B and Group C. A series of activities, presented in Table 2, were performed for the different groups, respecting the stipulated activity durations and sequence.

Table 2

Activities performed at the Solomon-Four Group experiments
Activity	Length (min)	Description	Group A	Group B	Group C
Introduction	20	AM capabilities variation across AM technologies, machines, and geometries. Introduction to the experiment´s tasks.	Introduction to document 1, pre-test and test tasks, and design platform.	Introduction to document 1 and pre-test and test tasks.	Introduction to document 1 and test tasks.
Read Tasks documentation	5–8	Participants read the experiment´s documentation on their own. They have 5 minutes that can be extended to 8 if they needed (in case of language difficulties, for example).	Read documents 1, 2 and 3	Read documents 1, 2 and 3	Read documents 1, and 3
Pre-test	20	Participants are presented with the pre-test redesign assessment activity, where they are tasked with voicing their concerns about the manufacturability and test ability of the product redesigned for AM.	Flow connector or heat exchanger (depending on iteration)	Flow connector or heat exchanger (depending on iteration)	-
Introduction to test	5	Transition between pre-test tasks and test tasks. Participants are given 5 minutes to read the documentation for the test	Read document 3, about the flow connector or heat exchanger (depending on iteration)	Read document 3, about the flow connector or heat exchanger (depending on iteration)	-
Digital platform presentation	10	Participants are introduced the digital design platform. On the platform, both the flow connector and the heat exchanger are already modelled and connected with the test activities PERT diagram.	Emphasis on the models of the product used for the test	Emphasis on the models of the product used for the test	Emphasis on the models of the product used for the test
Test	20	Participants are presented with the test redesign assessment activity, where they are tasked with using the digital design platform for voicing their concerns about the manufacturability and test ability of the product redesigned for AM.	Heat exchanger or flow connector (depending on iteration)	Heat exchanger or flow connector (depending on iteration)	Heat exchanger or flow connector (depending on iteration)

Three different documents are handled to groups A and B:

Document 1, AM capabilities and limitations: The experiments are based on metal laser powder bed fusion (LPBF) AM technologies; this document summarizes what can and cannot be manufactured with these technologies. Among others, minimum manufacturable wall thickness or threshold surface angle for requiring supports structures.
Document 2, pre-test documentation: The conventional geometry, materials, dimensions, and manufacturing processes of the product used for the pre-test; a table with the required product validation activities, the elements required for each activity, the activity duration, and the risk of not performing the activity; a PERT diagram displaying the validation activities sequence in the context of the whole fluid management system; the product redesigned for AM. Depending on the experiment iteration, document 2 can be about the flow connector (Iteration 1) or the heat exchanger (Iteration 2).
Document 3, test documentation: The conventional geometry, materials, dimensions, and manufacturing processes of the product used for the test; a table with the required product validation activities, the elements required for each activity, the activity duration, and the risk of not performing the activity; a PERT diagram displaying the validation activities sequence in the context of the whole fluid management system; the product redesigned for AM. Depending on the experiment iteration, document 2 can be about the flow connector (Iteration 2) or the heat exchanger (Iteration 1)

Groups C received only documents 1 and 3, as they do not have a pre-test.

The purpose of the documentation is to mimic real design processes and the information that is collected and available at conceptual stages. In Fig. 6, an extract of the PERT diagram (6.a) and validation activities table (6.b) are presented.

3.4. Variables of interest

During the study, participants are asked to voice their concerns related to:

(1) design manufacturability (weather physical features in the product are manufacturable with LPBF technologies or not).

(2) manufacturing processes (process parameters, manufacturing speed, manufacturing costs, post processing activities, “hybrid manufacturing”).

(3) test ability (ability to test the product or its physical features, how the AM design would affect test schedules in terms of duration and costs) (Valfre, 2012).

These concerns are “matters that cause feelings of unease, uncertainty, or apprehension” (MerriamWebster, 2021) to practitioners about technology capabilities (what can and cannot be achieved with AM) and the respective VVT activities.

What is measured in the experiments is the number of concerns (uncertainties) and the number of proposed design changes:

Number of concerns about manufacturability (Nm): weather physical features in the product are manufacturable or not.
Number of concerns about manufacturing processes (Nmp): process parameters, manufacturing speed, manufacturing costs, post processing activities, “hybrid manufacturing.
Number of concerns about test ability (Nt): ability to test the product or its physical features, how the AM design would affect test schedules in terms of duration and costs.
Number of proposed measures to address design manufacturing concerns, includes manufacturability and manufacturing processes (NMm).
Number of proposed measures to address test concerns (NMt).

The success criteria (Blessing et al., 1998) utilized to probe H1 and H2 are a higher number of concerns and proposed measures observed with the use of the design platform than without it.

In Table 3 the measured variables from the experiment are presented, categorized according to group and experiment iteration. The specification “Method” or “No method” refers to the use or not of the proposed design platform and constraints replacement procedure.

In Table 4, the differences between the two case studies, flow connector and heat exchanger, are presented. In Table 4, the influence of the pre-tests is also highlighted, as the results are categorized as “No practice” (no pre-test) or “With practice” (with pre-test).

Table 3

Measured variables from the Solomon Four-Groups experiments.
Iteration	Group		Nm	Nmp	Nt	NMm	NMt	Risk mentioned in the discussion
1	A	Pretest – No method	4	0	1	0	1	No
	A	Test - Method	7	1	2	0	5	Yes
	B	Pretest – No method	6	4	0	6	1	No
	B	Test – No method	5	1	2	1	4	No
	C	Pretest	-	-	-	-	-	-
	C	Test - Method	9	8	5	2	3	Yes
2	A	Pretest – No method	4	1	0	2	2	Yes
	A	Test - Method	9	4	1	1	4	Yes
	B	Pretest – No method	4	3	1	3	0	No
	B	Test – No method	3	1	0	2	0	No
	C	Pretest	-	-	-	-	-	-
	C	Test - Method	10	6	3	6	6	Yes

Table 4

Comparison of flow connector and heat exchanger case studies.
		Flow connector		Heat exchanger
		No Method	Method	No Method	Method
Nm	No Practice	4	10	4	9
	No Practice	6	-	4	-
	With Practice
Nmp	No Practice	0	6	1	8
	No Practice	1	-	3	-
	With Practice	1	4	1	1
Nt	No Practice	1	3	0	5
	No Practice	2	-	1	-
	With Practice	0	1	2	2
NMm	No Practice	0	5	2	2
	No Practice	0	-	3	-
	With Practice	2	1	1	0
NMt	No Practice	1	6	2	3
	No Practice	5	-	0	-
	Practice	0	4	4	5

From Tables 3 and 4, four results can be obtained:

The measured variables do not depend on the case study, the flow connector and the heat exchanger obtained similar results for every variable.
The pre-test activities do not seem to have a positive impact on the test results. Moreover, in the pre-test seems to have a negative effect on the amount of the number of concerns identified and the number of proposed measures to address uncertainties.
For both case studies, the number of concerns identified and the number of proposed measures to deal with uncertainty increased with the use of the design platform.
In every instance where the design platform was implemented practitioners held discussions about activity risk and used the provided risk assessments to propose measures to deal with test and manufacturing uncertainties. Only one group (It2 GA) included risk in their discussions without the support from the design platform.

In Table 5, the identified manufacturing and test concerns and measures are presented in relation to the group that identified them. Concerns explicitly included in the design platform as concerns for machining technologies are highlighted with an asterisk (*).

Table 5

Identified manufacturing and test concerns and measures are presented in relation to the group that identified them.
	Iteration 1
	A		B		C	A		B		C
Manufacturing concern	FC	HE	FC	HE	HE	HE	FC	HE	FC	FC
Max. manufacturing volume *		x			x		x			x
Min. manufacturable radius *		x					x			x
Min. manufacturable hole diameter*					x		x			x
Geometry changes between adjacent sections			x							x
Threshold angle for support structures	x	x	x	x			x	x		x
Max. diameter for horizontal holes without support			x		x
Part geometry to enable removal of support			x
Max. manufacturable aspect ratio *		x			x		x			x
Min. manufacturable wall thickness *	x	x	x		x		x	x	x	x
Max. length, diameter, wall thickness ratio *		x	x		x		x
Min. achievable surface roughness *	x	x		x			x			x
Surface quality improvement			x
Max. manufacturing speed to avoid overheating *		x			x		x			x
Process parameters to avoid deformations			x							x
Dimensional accuracy	x						x	x	x
Recoating blade problems (blade crush)			x					x
Part removal from building plate			x
Influence of manufacturing orientation				x	x	x	x	x	x	x
Anker the part to the building plate				x
Min. feature size to enable removal of powder				x		x
Mechanical properties homogeneity					x					x
Min. feature size dependency on the material					x
Influence of layer thickness					x
Influence of idle time between layers					x
Influence of manufacturing atmosphere					x
Influence of manufacturing temperatures					x
Influence of number of lasers					x
Number of parts on the building plate					x		x
Influence of location on the building plate					x	x
Wall thickness separation between internal channels						x
Hybrid manufacturing (machining + AM)						x		x	x
Defects generation and overheating depending on material										x
Defects generation depending on manufacturing parameters										x
Test concerns
Influence of design on the test schedule	x
Fit in the test chamber *		x					x			x
Cost of changing a test		x
Inspection of internal geometries				x
CT scan penetration depth				x	x
Component criticality					x
Need for stress relief before plate removal					x
Test of variable mechanical properties					x
CT scan resolution					x
Influence of material on CT usefulness										x
Affordability of CT scans										x
Ultrasonic testing										x
Measures to address manuf. concerns		b			b		b			b
Change holes geometry			x			x		x	x	x
Make bottom self-supporting			x
Smooth sharp corners			x					x		x
Integrate smart fixtures			x	x				x
Make angles between channels larger than 90						x
Separate inlet from outlet						x
Separate in several components									x
Downscale the component							x			x
Increase wall thickness										x
Increase scanning time										x
Increase layer thickness										x
Measures to address test concerns
Perform a mock-up test schedule	x
Change a test		x	x		x		x			x
Change the design		x								x
Downscale the component										x
Remove a test		x				x	x
Accept a longer test schedule		x
Change the schedule (test order)										x
Gather further information for test removal		x
Introduce a new test			x	x			x
Consult CT expert				x
Perform serial testing				x
Reduce testing by improving process monitoring				x	x
Ensure process repeatability					x
Test a couple of representative samples					x		x			x
Test several components at the same time for distributing the test activities and shortening the schedule										x

4.1. Qualitative assessments from participants

During the experiments, practitioners were asked to provide feedback from the digital platform and the design experience from the experiments. Moreover, after the study was finished, the results were compiled and presented to the practitioners on the research project to open a discussion about design support platforms and uncertainty identification.

Practitioners agreed on the fact that without the digital platform, it is difficult to understand and analyze the influence that design changes (in this case, from the introduction of AM) have on the test schedule.

This problem seems to be present in regular design scenarios, as practitioners mentioned ‘(AM) Sales pitches are about complexity for free and constraints reduction, but (the cost of testing) is not as frequently mentioned’ and ‘the platform puts a finger in a place we currently do not focus a lot, which is testing and validation and how we can change the validation phases’. Linking the test activities with the architecture model allows designers to include testing on the early design trade-off analysis and make design choices that will reduce the cost and duration of the test phases.

The FM-PERT connection has also highlighted some concerns that need to be addressed for an effective introduction of AM in industrial contexts. For critical components, validation is of extreme importance, however, due to the physical phenomena that takes place during AM processes, the validation activities of AM components become too expensive and time consuming. To reduce costs and foster AM industrialization, the focus must be set on ensuring manufacturing repeatability and validation. To smoothen the transition between current AM capabilities and fully repeatable AM processes, practitioners proposed improving the design platform with AI technologies:

‘The platform can be combined with AI technologies to feed the model with the likelihood of build success, it could be used to reduce risk. Which would be something that evolves and improves over time. So, you start with very high validation costs and over time (when the AI gains experience on manufacturing assessment) you reduce costs with AI’.

The reason why the digital design platform enables the inclusion of test phases during design decision making on conceptual stages is its visualization capability. Most of the practitioners mentioned that the platform helped them visualize product architecture and its influence on test activities ‘(the design platform) helps with visualization and (design) choices’. Moreover, the level of detail is appreciated, as it does not group too many things together which enables the platform evolution to fit new data and companies´ specific needs. The level of detail and the FM-PERT connection enables performing a sort of SWOT (Strengths, Weaknesses, Opportunities, Threats) analysis, ‘You can find opportunities to redesign quite early, you reduce reprinting processes (and costs) and redesign loops’. Overall, the design platform can serve to indicate where extra printing simulations (to gather missing data) might be necessary, instead of printing directly and risking the rework associated with failed manufacturing, ‘The connection with the PERT diagram is useful for experienced designers, they can easily see where the weaknesses and threats in the design are. Where do I need to focus? Where do I need to put more (simulation or test) resources?’.

The inclusion of the risk of not performing a test activity was appreciated, especially in the context of products where validation activities are traditionally time consuming and expensive, as is the case for the space industry. The benefit of including this risk on the FM and PERT diagrams is related with the possibility of considering not performing a test, if the risk of not performing this test is sufficiently low, and its implications on cost reduction ‘If you think about the component industry and all the quality assurance procedures that you can actually shave off your process…it will save you a lot of money immediately’. However, strong arguments are required to remove test activities and the design platform is not enough to make this type of decisions. Instead, it can point designers to open a conversation about this possibility and to gather the necessary data to make such decision ‘(The platform) is good to start a conversation about risk management and risk reduction. Maybe we can reduce the risk of some tests but we have to increase the part costs’.

In general, de design platform was recognized as a useful boundary object to foster communication among design teams and experts from different areas.

Practitioners indicated some of the design platform weaknesses as well.

On the current version, the platform does not include pictures of the product, which hinders the manufacturability assessment, especially for unexperienced practitioners.

Regarding the constraints replacement procedure, some practitioners indicated that is a practice that would benefit unexperienced practitioners mostly ‘The constraints replacement technique is mostly unnecessary for experienced practitioners’. Unexperienced designers can systematically go through the constraints for the previous implemented technology (in the case of the study, machining) and try finding their counterpart for AM. Through this procedure, they can find AM uncertainties previously unknown and learn about AM at the same time.

However, other practitioners indicated that that the design platform also benefits experienced designers as it guides their discussion (‘It points you out in the right direction’ ) and acts as a conversation starter, that fosters reflection and knowledge exchange. These statements are supported by the authors´ observations during the experiments.

Moreover, it was also observed that the design platform served as a knowledge transfer tool from experienced practitioners to unexperienced practitioners (‘I learned a lot today!’).

It was also suggested that more time was required to fully understand the design platform and make use of its full potential. Moreover, it was stated that the results using the platform might improve over time as designers gain experience and the platform is enhanced with new data ‘The more you use it, the better it gets. Is easy to adapt because the model is detailed’ .

In general, practitioners manifested that they were interested in trying the platform for their daily design activities.

The usefulness of a digital design platform for supporting design uncertainty identification and mitigation, was evaluated through an empirical study based on a Solomon Four Groups experiment design. Its usefulness was evaluated according to H1, useful for uncertainty identification and modelling, and H2, useful for transforming uncertainties into certainties.

The results suggest that the design platform enabled discussions about activity risk, and its use to propose measures to deal with uncertainties without compromising product quality while maintaining an affordable validation schedule. For example, the platform enables the identification of test activities with low risk and that can be removed or replaced with an inexpensive alternative.

However, strong arguments are required to remove test activities and the design platform is not enough to make this type of decisions. Instead, it can point designers to open a conversation about this possibility and to gather the necessary data or acquire the expertise to make such decision.

It was observed that implementing the design platform, the number of manufacturing and test concerns identified (uncertainties) and the number of proposed measures to deal with those uncertainties increased with the use of the design platform.

As the number of identified concerns increased with the implementation of the constraints replacement process, it can be concluded that this process is useful for identifying UU. Moreover, certain concerns were only identified through the constraints replacement method, such as “Max. manufacturing volume”, “Max. manufacturing speed to avoid overheating”, “Max. aspect ratio”, or “Fit in the test chamber”.

The relevance of identifying manufacturing and test related UU is highlighted, as non-functional requirements, such as manufacturability and test ability, are drivers of the design and decision-making process (Shankar et al., 2020).

However, a drawback from this technique is that UU, are discovered if their traditional technology counterpart is well modelled. In this sense, the usefulness of this process is related with the model´s depth, breadth, and fidelity (Haskins et al., 2015). The concerns “Min. feature size to enable removal of powder” and “Part geometry to enable removal of support structures” for example, are well-known AM concerns, however, they do not have a direct machining counterpart and were not mentioned by any group using the design platform.

These results resonate well with the work by. Roth et al. (2010) which state that the performance of past designs does not address all the sources of design uncertainty related with the introduction of new technologies. The main problem in this case is that not every constraint of a new technology can be found in a traditional technology counterpart. However, the design platform and the constraints replacement method itself serve as conversation openers, to initiate a discussion about technology uncertainties, and to promote knowledge exchange among practitioners within different disciplines. As proposed by Browning and Ramasesh (2015), an organization that is actively looking for UU is more likely to identify them and turn them into KU.

It was also observed during the experiments that many uncovered UU were not really UU, but aspects of the design, manufacturing process or testing that scaped the mind of the participants at the moment of the experiment. For instance, the concern “Max. manufacturing volume” is one of the more evident AM limitations, however, this concern was only mentioned by participants using the design platform and constraints replacement method. This observation was in line with literature (Ramasesh and Browning, 2014), in real design scenarios, UU are sometimes things that spaced the mind of practitioners or things that no one has bothered to find out. Moreover, they propose two types of UU, the knowable UU and the unknowable UU. The constraints replacement technique addresses the knowable UU.

Regarding UUs, there is a connection between the practitioners´ expertise and the number of UU they are able to identify with or without the design platform.

In group C from Iteration 1, one of the participants has 2 years of experience working exclusively with AM and design for AM, which enriched the discussion in this group. This group identified 17 manufacturing and manufacturing processes concerns, from the discussions enabled by the process of constraints replacement.

On the contrary, the participants of Group A from Iteration 1 where not that experienced in AM. Through the process of constraints replacement, this group identified eight manufacturing and manufacturing parameters concerns, seven of which were almost identical to their machining counterpart.

Workshops participants had different experience levels, and uncertainties that are unknown for a practitioner might be known for another. The design platform made explicit the knowledge of experienced practitioners, facilitating knowledge transfer to the less experienced.

The design platform and constraints replacement procedure support unexperienced designers, that otherwise, would not be able to identify as many uncertainties. Furthermore, for unexperienced practitioners the design platform serves as a training tool as well, as it was recognized as an efficient way to acquire and store information.

Furthermore, the platform´s customizable depth, breadth, and fidelity, allows for the adaptability and flexibility required when changes such as new technologies, new organizational processes or change of market focus, occur (Subrahmanianet al., 2003).

The results from the experiments suggest that while the platform supports uncertainty identification and modeling, it also supports the process of proposing affordable measures to deal with the uncertainties. While several measures to deal with uncertainty are experience dependent, the link between the FM and the PERT diagram highlights design and validation schedule changes that could reduce uncertainty or mitigate their effects in a cost and time efficient manner. For instance, only participants that did not use the platform proposed “Perform serial testing” as a measure to reduce uncertainties, however, this measure is time and resource intensive and can render a project unaffordable (Brice, 2011). Participants using the design platform had the possibility of simulating and observing the effect that these types of measures would have on the overall schedule affordability. Moreover, only participants using the platform proposed “Focus on process repeatability” to reduce the need for test activities performed on the product.

However, the platform is not intended to replace other design and uncertainty identification methods, rather, is intended to complement them. For example, practitioners stated that the platform would work better if it was connected to a CAD (physical) product representation that enables the analysis of a product´s geometry. These statements are in line with the work by authors such as McKoy et al. (2001) who sustain that graphical product representations are better than textual representations for engineering design idea generation processes.

Moreover, other techniques for UU identification, such as interviews, observations, workshops, scenarios, or prototypes (Sutcliffe and Sawyer, 2013), should not be left aside. They should be performed before or in parallel with the implementation of the design platform and complement its information. Designers should strive for a communication, and an alertness design culture that fosters the purposeful and systematic identification of UU (Ramasesh and Browning, 2014; Browning and Ramasesh, 2015).

The proposed design platform and other UU identification methods are a powerful complement for the multidisciplinary product development modelling strategies and performance simulations (Mavrik, 1999; Struck and Hensen, 2007; Ogaji et al., 2007; Goldberg et al., 2018) currently implemented for uncertainty assessment and reduction during conceptual stages.

In general, implementing the design platform, the number of uncertainties and proposed measures to deal with those uncertainties increased. However, the sample size of this study (which is its most relevant limitation) is too small and the obtained results cannot be generalized (Yin, 2003). Moreover, due to time and resources constrains, getting a randomized pool of experienced industrial practitioners able to participate in a one-and-a-half-hour experiment is very unlikely. In this situation, a non-random group of practitioners (from the available research project) and a non-random assignment of to the groups was necessary, this choice undermines the strength of the experiment as well.

Another limitation of this study is the artificial design setting. Practitioners from different companies were paired according to their availability and their previous experience with design and AM. In addition, one of the authors of this study was present during the experiments guiding the design analysis activities, which have likely impacted the results.

Moreover, experiment participants were unfamiliar with the type of product proposed for the case studies, as well as with each other. Team and case study familiarity is possibly another confounding variable in the study, research shows that the performance increases when team members are familiar with each other and trust each other (Arrow et al. 2000; Hargadon and Bechky, 2006).

Adding to the design setting artificial nature, the design intervals were restricted to 20 minutes each. On one side, some sessions needed to be cut short, it is possible that additional uncertainties and related mitigation measures would have been mentioned. On the other side, some other groups needed to “be forced” to talk during the 20 minutes, possibly due to the discomfort generated by a lack of perceived expertise on the unfamiliar case studies and a lack of familiarity with the design partner. Another consequence of the time-constrained design set up is that the pre-test seems to have a negative effect on the amount of the number of concerns identified and the number of proposed measures to address uncertainties during the test. One possible explanation that the authors find for this phenomenon is that the pre-test and test case studies we designed to foster the identification of similar concerns, to enable case study comparison. However, as both pre-test and test were performed within the same hour, having similar uncertainties, it is possible that some uncertainties that were mentioned on the pre-test were not mentioned on the test as an unconscious attempt to avoid redundancy.

It is evident that the artificial design setting affected the study results, however, literature shows that a large proportion of the research experiments performed for testing new tools and design methods are conducted in artificial settings and their results are still useful for industry and academia (Ellis and Dix, 2006).

Despite this study not presenting statistically relevant results, it was useful as a discussion enabler about design support platforms and a culture of uncertainty seeking. Moreover, the results of this study were used to obtain further funding that will enable the development of an improved version of the digital design platform that can be implemented in real industrial settings for further research purposes.

When introducing new technologies in product design, design uncertainties hinder the establishment of requirements during conceptual stages. Ill-defined design requirements can lead to expensive and time-consuming redesign loops.

Uncertainties can be known (information designers know is missing), or unknown (information designers do not know is missing). Unknown uncertainties present a big threat to a product development budget and schedule, as they are difficult to identify especially during early design phases.

In this article a digital design platform and a technique of constraints replacement are evaluated regarding their usefulness to identify new-technologies-related unknown uncertainties and measures to deal with them.

The platform and method are evaluated with a Solomon Four-Groups design study involving 12 experienced industrial practitioners and featuring the design of satellite components for additive manufacturing. The results of the study suggest that the proposed platform and associated design methods are useful for identifying uncertainties and for proposing measures to address them. Moreover, it servs as a conversation starter, to initiate discussions about technology uncertainties, fostering collaboration and knowledge transfer among practitioners with different expertise and from different disciplines. The results further highlight the need of making uncertainties seeking a common practice.

Despite the small sample size and the artificial design setting where the studies were performed, the results and conclusions obtained are useful as they invite a discussion about design support platforms and a culture of uncertainty seeking.

Acknowledgments

-Left out while in review

Arrow, Holly; Joseph E. McGrath; and Jennifer L. Berdahl (2000). Small groups as complex systems: formation, coordination, development, and adaption. Thousand Oaks, CA: Sage Publications.
Blessing, Lucienne T.; A. M. A. R. E. S. H. Chakrabarti; and K. M. Wallace (1998). An overview of descriptive studies in relation to a general design research methodology. Designers, 42-56.
Borgue Olivia; Jakob Müller; Alexander Leicht; Massimo Panarotto; and Ola Isaksson (2019). Constraint replacement-based design for additive manufacturing of satellite components: Ensuring design manufacturability through tailored test artefacts. Aerospace vol. 6, no. 11, pp. 124.
Brice, Craig (2011). Unintended consequences: How qualification constrains innovation. Proceedings of the 1st World Congress on Integrated Computational Materials Engineering (ICME), pp. 241. John Wiley & Sons.
Boujut Jean-François; and Eric Blanco (2003). Intermediary objects as a means to foster co-operation in engineering design. Computer Supported Cooperative Work (CSCW), vol. 12, no. 2, pp. 205-219.
Browning, Tyson R.; and Ranga Ramasesh (2015). Reducing unwelcome surprises in project management. MIT Sloan Management Review, vol. 56(3), pp. 53-62.
Chigbu, Uchendu Eugene (2019). Visually hypothesising in scientific paper writing: Confirming and refuting qualitative research hypotheses using diagrams. Publications, vol. 7(1), pp. 22.
Claesson Anders (2006). A configurable component framework supporting platform-based product development. PhD dissertation. Chalmers University of Technology: Dept. of Industrial and Materials Science.
Diegel, Olaf; Axel Nordin; and Damien Motte (2019). A Practical Guide to Design for Additive Manufacturing. Springer Singapore.
Dordlofva, Christo (2020). A Design for Qualification Framework for the Development of Additive Manufacturing Components—A Case Study from the Space Industry. Aerospace, vol. 7(3), pp. 25.
Dordlofva, Christo; and Peter Törlind (2020). Evaluating design uncertainties in additive manufacturing using design artefacts: examples from space industry. Design Science, vol. 6.
Ellis, Geoffrey; and Alan Dix (2006). An explorative analysis of user evaluation studies in information visualisation. Proceedings of the 2006 AVI workshop on Beyond time and errors: novel evaluation methods for information visualization, pp. 1-7. ACM, Venice, Italy.
Elverum, Christer W.; and Torgeir Welon (2015). On the use of directional and incremental prototyping in the development of high novelty products: Two case studies in the automotive industry. Journal of Engineering and Technology Management, vol. 38, pp. 71-88; https://doi.or g/10.1016/j.jengtecman.2015.09.003.
Frey, Daniel D.; and Clive L. Dym (2006). Validation of design methods: lessons from medicine. Research in Engineering Design, vol. 17, no. 1 (2006), pp. 45-57.
Jensen, Matilde B.; Christer W. Elverum; and Martin Steinert (2017). Eliciting unknown unknowns with prototypes: Introducing prototrials and prototrial-driven cultures. Design Studies, vol. 49, pp. 1-31.
Kleinsmann, Maaike; Fleur Deken: Andy Dong; and Kristina Lauche (2012). Development of design collaboration skills. Journal of Engineering Design, vol. 23, no. 7: pp. 485-506.
Kumke, Martin; Hagen Watschke; and Thomas Vietor (2016). A new methodological framework for design for additive manufacturing. Virtual and Physical Prototyping, vol. 11, no. 1, pp. 3–19.https://doi.org/10.1080/17452759.2016.1139377.
Goldberg, Chana; Devaiah Nalianda; V Sethi; Pericles Pilidis; Riti Singh; and Konstantinos Kyprianidis (2018). Assessment of an energy-efficient aircraft concept from a techno-economic perspective. Applied Energy, vol. 221, pp. 229-238.
Hargadon, Andrew B.; and Beth A. Bechky (2006). When collections of creatives become creative collectives: a field study of problem solving at work. Organization Science, vol. 17(4): pp. 484–500.
Henricksen, Karen; and Jadwiga Indulska (2004). Modelling and using imperfect context information. IEEE Annual Conference on Pervasive Computing and Communications Workshops, 2004. Proceedings of the Second, pp. 33-37. IEEE.
Mai, Nu Nu; Yoshi Takahashi; and Mon Mon Oo (2020). Testing the Effectiveness of Transfer Interventions Using Solomon Four-Group Designs. Education Sciences, vol. 10(4), pp. 92.
MerriamWebster (2021) Dictionary-Concern, Merriam Webster. https://www.merriam-webster.com/dictionary/concern#:~:text=Definition%20of%20concern%20%28Entry%202%20of%202%29%201,The%20actor%27s%20sudden%20collapse%20on%20stage%20caused%20concern. Accessed 27 May 2021.
Mavris, Dimitri N.; Oliver Bandte; and Daniel A. DeLaurentis (1999). Robust design simulation: a probabilistic approach to multidisciplinary design. Journal of Aircraft, vol. 36(1), pp. 298-307.
McKoy, Felicia L.; Noé Vargas-Hernández: Joshua D. Summers; and Jami J. Shah (2001) Influence of design representation on effectiveness of idea generation. International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, vol. 80258, pp. 39-48. American Society of Mechanical Engineers.
Ogaji, S. O. T.; Pericles Pilidis; and Richard Hales (2007). TERA-a tool for aero-engine modelling and management.
Palinkas, Lawrence A.; Sarah M. Horwitz; Carla A. Green; Jennifer P. Wisdom; Naihua Duan; and Kimberly Hoagwood (2015). Purposeful sampling for qualitative data collection and analysis in mixed method implementation research. Administration and policy in mental health and mental health services research, vol. 42(5), pp. 533-544.
Passemard, Dan; and Brian H. Kleiner (2000). Competitive advantage in global industries. Management Research News.
Pedersen, Kjartan; Jan Emblemsvåg; Reid Bailey; Janet K. Allen; and Farrokh Mistree (2000). Validating design methods and research: the validation square. In Proceedings of the International Design Engineering Technical Conferences and Computers and Information in Engineering Conference, vol. 35142, pp. 379-390. American Society of Mechanical Engineers.
Porter, Michael E. (1985). Technology and competitive advantage. Journal of business strategy.
Raydugin, Yuri (2012). Quantifying unknown unknowns in an oil and gas capital project. International Journal of Risk and Contingency Management (IJRCM), vol. 1(2), pp. 29-42.
Ramasesh, Ranga V.; and Tyson R. Browning (2014). A conceptual framework for tackling knowable unknown unknowns in project management. Journal of operations management, vol. 32(4), pp. 190-204.
Sawilowsky, Shlomo; D. Lynn Kelley; R. Clifford Blair; and Barry S. Markman (1994). Meta-analysis and the Solomon four-group design. The Journal of Experimental Education, vol. 62(4), pp. 361-376.
Seepersad, Carolyn Conner; Jared Allison; and Conner Sharpe (2017). The need for effective design guides in additive manufacturing. Proceedings of the 21st International Conference on Engineering Design (ICED 17), vol. 5: Design for X, Design to X, Vancouver, Canada, pp. 309-316.
Shankar, Prabhu, Beshoy Morkos, Darshan Yadav, and Joshua D. Summers (2020). Towards the formalization of non-functional requirements in conceptual design. Research in Engineering Design, vol. 31, no. 4, pp. 449-469.
Stroud, Dick (2002). The impact of the new technologies—Permanent paradigm shift or transitory aberration?. Interactive Marketing, vol. 4(2), pp. 144-155.
Struck, Christian; and Jan Hensen (2007). On supporting design decisions in conceptual design addressing specification uncertainties using performance simulation. Proceedings of the 10th IBPSA Building Simulation Conference, pp. 3-5.
Subrahmanian Eswaran; Ira Monarch; Suresh Konda; Helen Granger; Russ Milliken, and Arthur Westerberg (2003). Boundary objects and prototypes at the interfaces of engineering design. Computer Supported Cooperative Work (CSCW), vol. 12, no. 2, pp. 185-203.
Sutcliffe, Alistair; and Pete Sawyer (2013). Requirements elicitation: Towards the unknown unknowns. 21st IEEE International Requirements Engineering Conference, pp. 92-104. IEEE.
Tahera, Khadija; David C. Wynn; Chris Earl; and Claudia M. Eckert (2019). Testing in the incremental design and development of complex products. Research in Engineering Design vol. 30, no. 2291-316.
Thompson, Mary Kathryn; Giovanni Moroni; Tom Vaneker; Georges Fadel; Ian Campbell, Ian Gibson; Alain Bernard et al. (2016). Design for Additive Manufacturing: Trends, opportunities, considerations, and constraints. CIRP Annals - Manufacturing Technology vol. 65, pp. 737-760; https://doi.org/10.1016/j.cirp.2016.05.004.
Trochim William (2021). Experimental design. https://conjointly.com/kb/experimental-design/. Accessed 27 May 2021.
Valfre James (2012). Testability modeling usage in design-for-test and product lifecycle cost reduction. 2012 IEEE AUTOTESTCON Proceedings, pp. 39-41. IEEE.
Yin Robert (1994). Case study research: design and methods. Sage Publications, Inc ISBN 0 761, no. 92553, pp. 8.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Reducing design uncertainty through model-based collaborative design methods when introducing new technologies: A Solomon four-groups empirical design study

Status:

Version 1

Abstract

Figures

1. Introduction

2. Background

2.1. Uncertainties during technology introduction

2.2. A multidisciplinary design platform to model uncertainty and identify unknown unknowns

2.3. Experimental design methods for evaluating the usefulness of a method

3. Method

3.1. Solomon-four group design experiment

3.2. Case studies for pre-tests and tests

3.3. Activities performed during the experiments

3.4. Variables of interest

4. Results

4.1. Qualitative assessments from participants

5. Discussion

6. Conclusion

Declarations

Acknowledgments

References

Additional Declarations

Status:

Version 1