Applying Creativity to Generating a Concept

Systems Architecture · Chapter 12

Aykut C. Satici

Generating Concepts from Goals

Stakeholder analysis and goal writing (Ch. 11) reduce ambiguity. Developing the system concept is fundamentally a creative process
The architect must be prepared to architect “up” (toward stakeholder needs) and “down” (toward architecture and operations) from the concept: not just receive a concept from above
In Chapter 7 we analyzed the parts of a concept. Here we use that analysis to structure how a new concept gets created: with emphasis on applying creativity, worked through a Hybrid Car example

Place this chapter in the arc of Part 3: Ch. 11 reduced ambiguity (fuzzy needs → prioritized goals), disciplined analysis. This chapter is different in character: generating the concept is a creative act: a deliberate shift in mode
Second bullet corrects a common misconception: concept isn’t simply handed down from stakeholders for the architect to receive. Architect works up (stakeholder needs, value goals) and down (architecture, operations): concept sits in the middle, not atop a one-way chain
Preview Figure 12.1 (next slide): concept is the pivot point of the whole architecting process
Tie to Ch. 7: anatomy of a concept = mapping from function to form, To-By-Using framework. This chapter asks the generative question: given that structure, how do we actually create one?
Hybrid Car is the running worked example, continuing directly from Ch. 11’s stakeholder/goal work

Suggested answer: Start from the solution-neutral intent and explore multiple ways to specialize the operand and process, then identify enabling forms. Expand rich concepts into fragments, combine compatible fragments, and evaluate the integrated alternatives against the goals. For moving a vehicle, this means considering different energy-conversion and storage arrangements rather than assuming an engine or battery first. Separate generating possibilities from prematurely ranking them.

Figure 12.1 · Three Themes in Architecting

Concept sits at the center: stakeholder needs and value goals flow down into it (reducing ambiguity); architecture and operations flow out of it (managing complexity); creativity is applied at the concept itself. Source: Crawley, Cameron & Selva (2016), Fig. 12.1.

Visual anchor for the whole Part 3 structure, not just this chapter: concept sits literally in the middle
Above: stakeholder needs/value goals flow down → “reducing ambiguity” (Ch. 9-11)
Below: architecture/operations flow out → “managing complexity” (Ch. 13, next)
At concept itself: applying creativity: this chapter’s contribution, hinge between the other two themes
This figure is effectively Part 3’s table of contents (reduce ambiguity, apply creativity, manage complexity), concept as connective tissue. Have students locate “where we are” on this diagram at the start of each upcoming chapter: repeating orientation device

Discussion: Why is concept generation between ambiguity reduction and complexity management?

Suggested answer: Goals define what outcomes matter; concept generation explores how to achieve them; decomposition develops the chosen approach. Feedback between all three can revise earlier assumptions.

12.2 Applying Creativity to Concept

What Is Creativity?

Creativity must result in a novel output: on this, most agree. Less agreement: must it be intentional? Must it have influence or impact?
We take the position that creativity must be intentional: accidentally spilling paint on a canvas and discarding it isn’t creative: but creativity should not be defined by impact, since a focus on impact too early can restrict ideation

Generate a range of concepts before selecting among them. Alex Osborn expressed this approach as “quantity breeds quality.”

The book stakes out a precise position on a contested term rather than leaving it vague. Everyone agrees creativity must be novel. Disagreement: must it be intentional? judged by impact?
Intentionality: book’s position: creativity must be intentional. Memorable counter-example: spilling paint accidentally, then discarding it: not creative, technically novel but no deliberate effort. Intent doesn’t require a known/linear process, just deliberately applied effort
Impact: more counter-intuitive, explain why: creativity should not be defined by eventual influence/impact: judging by impact too early is premature filtering that kills ideation. Suppresses raw, half-formed ideas that might mature later. Impact assessment is the job of metrics applied afterward (Pugh matrices, later this chapter), not an ideation-stage gate
Osborn’s phrase “quantity breeds quality” encourages generating alternatives before screening them.
Figure 12.2 shows expansion followed by selection. Figure 12.10 returns to this sequence.

Suggested answer: Within the book’s chosen definition, creativity involves deliberate effort producing something novel; it need not already have demonstrated market or social impact. An accident can supply a stimulus, but recognizing, developing, and purposefully applying it is the creative act. Other definitions exist, so distinguish this course convention from a universal philosophical verdict. Evaluate usefulness and feasibility during concept assessment without demanding proven impact from every early idea.

Figure 12.2 · Applying Intentional Creativity

Applying intentional creativity to concept means expanding the number of concepts under consideration, then winnowing the list according to “fit against goals.” Source: Crawley, Cameron & Selva (2016), Fig. 12.2.

Unstructured vs. Structured Creativity

Unstructured Creativity

The more prevalent approach: brainstorming, blue-sky ideas, free association. Focuses on ideating without prejudices from previous experience: forming new pathways through the concept-space (de Bono).

Structured Creativity

Holds that problem analysis can help solution synthesis. Creative thinking is not different from ordinary problem solving: creativity can be stimulated through analysis.

Tip

This course uses structured methods alongside brainstorming to generate and compare alternatives.

Two broad schools of thought on ideation: rest of the chapter’s tools sort cleanly into one or the other
Unstructured creativity: brainstorming, blue-sky, free association: rooted in unconscious processing, an “Aha!” that seems to arise from nowhere. Edward de Bono: mind is physically constrained by existing thought-channels; unstructured techniques force new pathways (e.g., open a book to a random page, point blindly, force a connection to the problem)
- Explain: some hold creativity is by definition unconscious, must occur without warning: if plannable, not really creative. Book calls this feasible but “somewhat counterproductive”
Structured creativity: analyzing the problem helps synthesize the solution: creative thinking isn’t different from ordinary problem-solving, can be stimulated by structured analysis
- Famous “creative thinkers” (Einstein, Picasso, da Vinci, Maya Angelou) show strong divergent thinking: counterpoint: this capacity can be built through structure, not just innate/unconscious
Book’s considered position: repeat to students, explains the shape of the rest of the chapter: both necessary, but leans structured: would rather succeed with an architecture chosen through traceable reasoning than arrived at by luck
Same sentiment as Ch. 1’s “structured creativity is better than unstructured,” now fully grounded methodologically

Discussion: When would you use structured methods after brainstorming?

Suggested answer: When ideas are concentrated in familiar areas or missing entire mechanisms. A framework prompts neglected alternatives while still requiring judgment about usefulness and feasibility.

Figure 12.3 · Component Recombination

A frequent theme in structured creativity: decompose the problem into pieces (here, the transporting concept’s three internal functions), with more than one form-choice available per piece: choosing one option per row (shaded) yields distinct concepts, as in Ch. 7’s morphological matrix. Source: Crawley, Cameron & Selva (2016), Fig. 12.3.

First concrete structured-creativity technique: component recombination
Diagram’s example: transporting → three internal functions (lifting, propelling, guiding), each with more than one form-choice (wheels, propeller, etc.) per row
Synthesize a concept by selecting one choice per row; different row-selections → different concepts
Diagram deliberately vague about form–function assignments vs. specific functions for a solution-neutral function: intentional, same recombination logic applies either way
Same logic as Ch. 7’s morphological matrix (propeller for lifting+propelling+guiding = helicopter). Recurs in Ch. 14’s decision-support tools: a recurring technique, not one-off

Discussion: Does recombining familiar components preclude creativity?

Suggested answer: No. New relationships can produce a novel system-level approach. The combination must be coherent and should be evaluated for its resulting function, not only its unfamiliar appearance.

Completeness Frameworks and TRIZ

Completeness frameworks use lists to stimulate ideation: e.g., cataloging forms of energy (linear kinetic, rotational kinetic, potential, chemical…) and asking what a concept using each would look like
De Bono’s Six Hats is a team-based completeness framework: six roles (Managing, Information, Emotions, Discernment, Optimistic Response, Creativity) that force a more holistic evaluation of a problem
TRIZ (Theory of Inventive Problem Solving), developed by Genrich Altshuller from 40,000 patent abstracts, resolves apparent contradictions (“a faster train needs a more powerful engine, but a more powerful engine is heavier”) using 40 inventive principles: e.g., Mechanics Substitution: replace a mechanical means with an optical, acoustic, magnetic, or electromagnetic one

Second family: completeness frameworks: pre-built lists systematically stimulate ideation, rather than recombining decomposed pieces
Energy-forms example: checklist of energy forms (linear kinetic, rotational kinetic, potential, chemical…): for each, ask “what would a concept using this look like?” Forces coverage a free-associative brainstorm might miss
- Concrete: hybrids mostly store energy chemically (batteries), but rotational kinetic exists too: flywheels in Porsche’s GT3RS hybrid and the 2009 Williams F1 car
De Bono’s Six Hats: team-based version: six roles (Managing, Information, Emotions, Discernment, Optimistic Response, Creativity), cycling the team through different reasoning modes for a more holistic evaluation
- Be honest: not proven orthogonal cognitively, but still a useful, practical decomposition of team roles
TRIZ (Theory of Inventive Problem Solving): most time of the three, real engineering pedigree
- Genrich Altshuller, Soviet Navy, reviewed ~40,000 patent abstracts → 40 general principles of invention
- Starts by naming apparent contradictions: “a faster train needs a more powerful engine, but a more powerful engine is heavier”
- Example principle: Mechanics Substitution: replace a mechanical means with a sensory one (optical, acoustic, magnetic, electromagnetic): e.g., diesel locomotive → electric locomotive or linear induction motor
- → Caveat: principles don’t guarantee the same weight or any specific property: they give a direction/concept to explore, not a guaranteed optimized outcome

Suggested answer: For recovering vehicle braking energy, an energy-form checklist suggests electrical storage in a battery, rotational kinetic storage in a flywheel, or pressurized-fluid storage in a hydraulic accumulator. Each is a concept fragment that needs an appropriate conversion path and feasibility analysis. A framework can reveal alternatives missed by habit; it does not prove completeness or make every generated alternative useful. For TRIZ, describe the conflicting requirements first, then test whether a proposed principle actually resolves the conflict in the full system.

Identifying Concept

Recall from Chapter 7: concept is a vision, idea, notion, or mental image that maps function to form: it embodies a principle of function and operation, and includes an abstraction of form
Concept selection determines the proposed operating principle and much of the vocabulary used in later design
“The concept rationalizes the structure of the architecture.”: Steve Imrich. Concept is not a product attribute; it is a mapping from one attribute (function) to another (form): separate from, and prior to, the architecture itself

Callback-and-consolidate before the four-step framework
Restate Ch. 7’s definition precisely: everything downstream depends on it: concept = vision/idea/notion/mental image mapping function to form, embodies a principle of function and operation, includes an abstraction of form (not full detail)
Selecting a concept affects later design choices and establishes the terminology used by the teams developing it.
Steve Imrich (civil architect: note the domain-crossing citation): “the concept rationalizes the structure of the architecture.” Rationalizes = gives the architecture its underlying logic, the reason it’s organized this way and not another
Precise, easy to get backwards: concept is not a product attribute: it’s a mapping from one attribute (function) to another (form). Separate from, and logically prior to, the architecture itself

Discussion: How do you tell a concept from a desired benefit?

Suggested answer: The benefit states what is wanted; the concept identifies a specific process and enabling form that could provide it. Keep both explicit to compare alternative means fairly.

Figure 12.4 · Representation of Concept

Concept sits along a diagonal between pure function and pure form: neither fully abstract nor fully concrete
In Chapter 11 we analyzed concepts in OPM, completing the To-By-Using framework
For complex systems, OPM may only work at the first level: domain-specific language and methods quickly become necessary

Form

↗

Concept

Function

Diagonal-arrow diagram: compact but informative: concept sits on a diagonal from pure function (bottom) to pure form (top), deliberately neither fully abstract nor fully concrete
Positions concept exactly where the Steve Imrich quote located it: a bridge between function and form
Connect to Ch. 11: OPM analyzed concepts there, completing To-By-Using (To = value, By = process, Using = instrument/form) formally in OPM notation
Scope-setting caveat, ties to the course’s representation-tool policy: for complex systems, OPM may only stay tractable at the first level: domain-specific language/methods become necessary quickly. Remind students: OPM is our primary tool, but real complex problems eventually need domain expertise on top

Discussion: What information is lost if the concept is drawn only as hardware boxes?

Suggested answer: The operand, process, and intended transformation may disappear. An OPM-oriented explanation restores how the form enables function rather than leaving behavior to inference.

Box 12.1 · Principle of Creativity

“Imagination is more important than knowledge.”: Albert Einstein

“Change takes place through the struggle of opposites.”: Vladimir Lenin

“When you are face to face with a difficulty, you are up against a discovery.”: Lord Kelvin

“Creativity is bred by creating a gap between current reality and the vision for the system.”: Peter Senge

Architectural creativity resolves tensions among system goals. If a feasible concept cannot meet them, revisit the goals.

Reframes what “creativity in architecture” actually means
Central claim: any interesting problem has essential tensions among goals (fuel efficiency vs. power, cost vs. capability: familiar from Hybrid Car goal-setting in Ch. 11). Creativity = the process of resolving these tensions, not generating ideas in a vacuum
Three actionable practices to maximize creativity: treat all goals as tradable; remove organizational/cultural barriers suppressing ideas; deliberately consider the entire space of possibility, not an early comfortable subset
Einstein’s quote: imagination exceeding known facts
Doctorow’s headlights quote: a useful mental model: you never see further than your headlights, but you can make the whole trip that way. Comforting reframe: you don’t need to see the whole solution before generating concepts: just as far as your current headlights

Discussion: Does a structured creativity framework guarantee a good concept?

Suggested answer: No. It improves exploration and makes assumptions visible, but generated candidates still need functional coherence, feasibility checks, and assessment against stakeholder goals.

Box 12.1 · How to Create Room for Alternatives

Treat goals as tradable while exploring concepts; do not silently discard a goal to make one favorite design work
Remove organizational and cultural barriers that prevent people from proposing unfamiliar options
Search the full plausible space, including new associations between existing ideas
Iterate: a concept may resolve the tension, or evidence may show that the goals need revision

E. L. Doctorow’s driving-at-night analogy: progress is possible without seeing the entire route in advance.

A Four-Step Concept Framework

1. Develop the Concepts: start with the SPS and goals; identify solution-neutral operands/processes; apply creativity to specialize a specific operand/process/instrument; check against goals.

2. Expand and Develop Concept Fragments: decompose rich, multifunctional concepts to reveal internal functions; repeat Step 1 for each fragment.

3. Evolve and Refine Integrated Concepts: search systematically through fragments; combine combinatorially, with constraints.

4. Select a Few Integrated Concepts: apply backward (fit to goals) and forward (potential for good architecture) considerations.

Chapter’s central organizing structure: rest of the chapter (through 12.6) is a detailed worked example of these four steps on the Hybrid Car. Introduce all four as a roadmap first
Step 1: Develop the Concepts: start from SPS + Ch. 11 goals; identify solution-neutral operands/processes; apply creativity to specialize a specific operand/process/instrument; check against goals, reformulate if it fails
Step 2: Expand and Develop Concept Fragments: for rich/multifunctional concepts, decompose to reveal internal functions; repeat Step 1 for each fragment
Step 3: Evolve and Refine Integrated Concepts: search systematically for coverage, combine fragments combinatorially (subject to constraints) into integrated concepts
Step 4: Select a Few Integrated Concepts: backward consideration (fit to goals) + forward consideration (potential for good architecture); re-check against SPS/goals
Steps 1-2 = “ballooning” half of Figure 12.2’s diamond; Steps 3-4 = “narrowing” half: this framework operationalizes the diamond into four repeatable steps

Discussion: Why select several concepts rather than one immediately?

Suggested answer: Early evidence is incomplete, and later integration or evaluation may reverse the ranking. A small diverse set preserves useful alternatives without carrying every idea into detailed design.

Section 12.2 Summary

Creativity must be intentional, but should not be judged by impact during ideation: that comes later, from metrics
Unstructured creativity (brainstorming) and structured creativity (component recombination, completeness frameworks, TRIZ) are complementary, not competing, approaches
Concept is a mapping from function to form, developed through a four-step framework: develop, expand into fragments, evolve/refine integrated concepts, and select a few for further development

Review before shifting from theory (12.2) to the fully worked Hybrid Car example (12.3-12.6)
Students should be able to state: creativity must be intentional but not judged by impact during ideation (deferred to metrics afterward); unstructured and structured creativity are complementary, not rivals, though this course leans structured; concept = mapping from function to form via the four-step framework (develop, expand into fragments, evolve/refine integrated concepts, select a few)
Everything ahead is that same framework applied in full, concrete detail to one running example: the Hybrid Car

Discussion: What should a team produce before detailed concept scoring?

Suggested answer: A clear intent, distinguishable concepts, and enough integrated detail to compare them consistently. Scoring vague labels creates precision without a common basis.

12.3–12.5 Developing the Hybrid Car Concept

Step 1: Develop the Concept

Recall the Hybrid Car’s system problem statement from Chapter 11:

Provide our customers a product to transport them and their possessions inexpensively and in an environmentally sound manner, by allowing them to drive themselves, their passengers, and light cargo fuel-efficiently and with good handling characteristics, using a hybrid gas/electric car.

The word “hybrid” is itself a short form: it suppresses which two energy poles are combined. Historically, there was even a time when “hybrid” commonly meant hybrid steam/gas, not gas/electric

Begin the fully worked example by re-grounding: this SPS is the same To-By-Using statement from Ch. 11’s stakeholder/goal analysis, carried forward unchanged
Read aloud slowly: To (transport inexpensively, environmentally sound) = value delivered; By (drive themselves/passengers/light cargo, fuel-efficiently, good handling) = process; Using (hybrid gas/electric car) = instrument/form, already partially specialized
Historical seed for the closing case study: “hybrid” suppresses which two energy poles combine. Today assumes gas+electric, but historically “hybrid” often meant steam/gas: pays off in the case study on electric/steam/ICE architectural competition

Discussion: Why begin the hybrid example with moving the vehicle rather than a battery?

Suggested answer: Moving defines the functional intent. Starting with a battery fixes one form prematurely and can exclude other energy-storage or propulsion concepts.

Step 2: Expand the Concept: Propulsion

“Driving” includes several internal functions. Begin the decomposition with propulsion
Three overarching vehicle classifications, based on how many external energy sources the propulsion system depends on:

Monovalent

One external energy source (most cars today, including simple hybrids with one fuel source)

Bivalent

Two external energy sources (e.g., a plug-in hybrid: electricity and fuel)

Multivalent

Three or more (e.g., the Fiat Siena Tetrafuel: gasoline, ethanol blends, or CNG)

Step 2 applied for the first time: “driving” = Step 1’s specific process, flagged “rich in meaning” (bundles several internal functions): the trigger condition for Step 2. First fragment: propulsion
Three-way propulsion classification, based on number of external energy sources:
- Monovalent: one external source: most cars today, including simple hybrids (still only gasoline externally, battery is an internal secondary store)
- Bivalent: two external sources: plug-in hybrid (electricity from grid + gasoline)
- Multivalent: three or more: Fiat Siena Tetrafuel (gasoline, E20-E25 ethanol blends, pure ethanol E100, or CNG)
This classification matters directly later, when asking which category the Hybrid Car concept should specialize into

Discussion: What does considering several propulsion sources add to exploration?

Suggested answer: It reveals possible energy-conversion arrangements and combinations. Additional sources also introduce interfaces and control needs, so variety alone does not establish value.

Source Categories versus Concept Fragments

Classification is not yet a concept fragment. Monovalent, bivalent, and multivalent count external energy sources; none specifies which form performs a function.

Classification: a plug-in hybrid can take energy from fuel and the electrical grid
Concept fragment: a fuel converter drives the wheels, or drives a generator that powers an electric traction motor
Count external supplies for each candidate: a fuel-only HEV is monovalent; adding grid charging makes it bivalent

Figure 12.5 · Four Vehicle-Moving Concepts

The textbook compares steam, internal-combustion, battery-based, and fuel-cell pathways, including their primary energy sources and carriers. Source: Crawley, Cameron & Selva (2016), Fig. 12.5.

Dense, information-rich: spend time here. Trace left to right: primary sources (fossil fuels, biomass, renewables, nuclear) → energy carriers (liquid fuels, gaseous fuels, electricity) → four vehicle-moving pathways: steam, ICE, electric (HEV→PHEV→BEV), fuel-cell
Steam: dominated ~1790-1906, no successful modern commercialization: flag that we return to why at chapter’s end
ICE: dominant today
Battery-based: a spectrum: HEV (small assist) → PHEV (plug-in) → BEV (no combustion engine at all)
Fuel-cell: converts fuel chemically into electricity for electric motors, rather than burning directly
This is the “energy carrying/storage” fragment of “driving,” expanded per Step 2: a separate fragment from propulsion-architecture (monovalent/bivalent/multivalent) on the previous slide, both expanded independently

Discussion: Are four illustrated moving concepts an exhaustive vehicle design space?

Suggested answer: No. They organize selected possibilities under the chapter’s scope. Other concepts or combinations may exist, and each illustrated concept still needs integration and evaluation.

Seven Additional Concept Fragments

Beyond propulsion, seven internal functions capture the value added by hybrid systems:

Motor start-stop: shut off the engine at rest, restart on demand
Regenerative braking: capture braking energy that would otherwise be lost to friction/heat
Power boost: electric motor adds torque beyond what the engine alone delivers
Load level increase: engine drives a generator to recharge the battery

Electric driving: propel using only stored electric energy, engine decoupled
External battery charging: plug into the grid (differentiates PHEVs from other HEVs)
Gliding: decouple both engine and electric system, coast on gravity alone

Catalogs the remaining seven fragments of “driving” capturing hybrid-added value beyond conventional ICE: these get directly reused later to build integrated vehicle concepts
Motor start-stop: engine shuts off at a (traffic light), restarted via motor/starter-generator when power needed: in micro-hybrids without electric driving, restart in well under a second, triggered by brake release/clutch depression
Regenerative braking: braking energy normally lost to friction/heat captured by running the traction motor generatively, stored in the high-voltage battery
Power boost: electric motor adds torque beyond what the engine alone delivers (inclines, towing): especially effective from rest, since electric motors peak at zero RPM, exactly where combustion engines are weakest (Figure 12.7)
Load level increase: engine runs a generator to recharge the battery, shifting to a more efficient operating point when excess power capacity is available
Electric driving: stored electric energy alone powers the traction motor, engine fully decoupled (off, or generating instead of propelling): limited by battery storage capacity
External battery charging: differentiates plug-in hybrids from every other type: charging unit replenishes from the grid, not just onboard sources; opens possibilities for nighttime charging and vehicle-to-grid power flow
Gliding: decouple both engine and electric system, coast on gravity alone, avoid powertrain friction losses: conventional cars do this via neutral downhill; hybrids need rapid reconnection when conditions change

Discussion: Can all seven additional fragments be added independently?

Suggested answer: Not necessarily. They may share equipment, interact through control, or depend on operating conditions. Assess the integrated combination rather than summing isolated benefits blindly.

Figure 12.6 · Fuel Savings from Start-Stop and Braking

Idling reduction via motor start-stop saves ~5–7%; an optimized hybrid control strategy (combining load level increase, boosting, electric driving, gliding) saves a further ~5–9%; regenerative braking saves another ~5–9%: totaling ~15–25%. Source: Crawley, Cameron & Selva (2016), Fig. 12.6, after Ehsani, Emadi & Gao (2009).

Numbers quantify what the fragments are worth: makes the abstract discussion concrete
Baseline: conventional ICE. Motor start-stop alone (eliminate idling): saves ~5-7% city driving
- optimized control strategy (load level increase, boosting, electric driving, gliding combined): further ~5-9%
- regenerative braking: further ~5-9%
Combined: ~15-25% savings vs. conventional baseline: real, quantified backward-considerations for comparing candidates against Ch. 11’s fuel-efficiency goal
Subtlety: external battery charging is absent from this figure: it adds energy from outside the vehicle, not a cleverer use of onboard energy. These percentages represent full hybrids with limited (non-plug-in) electric driving, not plug-in hybrids (which replace a much larger fuel share by design)

Discussion: Why might braking-energy recovery save little on a particular journey?

Suggested answer: Its benefit depends on opportunities to decelerate and recover energy, along with conversion and storage limits. A steady journey can offer fewer opportunities than stop-and-go operation.

Figure 12.7 · Why Boosting Helps at Low Speed

The electric motor delivers its highest torque starting from rest and low RPM: exactly where a combustion engine is weakest. Boosting lets the electric motor supply torque during the acceleration phase the engine handles worst. Source: Crawley, Cameron & Selva (2016), Fig. 12.7.

The physical reason power-boost works, not just an assertion
Electric motor’s torque curve: flat and high from zero RPM, tapers at higher speeds
Otto-cycle combustion engine’s curve: opposite: low at rest/low RPM, peaks at higher RPM (~2000-2500 RPM)
Complementary shapes are the purpose: exactly where combustion is weakest (from rest, low RPM) is exactly where electric is strongest: boosting especially improves 0-100 km/h acceleration by using electric torque from rest, then letting combustion take over its favorable range
Nice concrete example of “emergence depends on structure” (Ch. 2) applied to powertrains: neither curve alone is as good as the combination, because weaknesses/strengths are complementary, not overlapping

Discussion: Why consider motor boosting together with engine sizing and control?

Suggested answer: Electric assistance can change the engine’s required operating envelope. Its value depends on when power is available and how the complete propulsion system is sized and managed.

Step 3: Evolve and Refine Integrated Concepts

The third step searches the space of concept fragments systematically, organized by the possible mappings between functions and forms
Key question: is the energy storage function an input to the vehicle-moving function (a shared form: parallel hybrid), or do they use separate forms (series hybrid)?

Tip

Vehicles that mix both modes are combined hybrid systems: they may split fuel-converter energy across series and parallel paths simultaneously (power-split hybrids), or switch between the two.

Step 3: evolve and refine integrated concepts, searching systematically through Step 2’s fragments
Book candidly excerpts a full combinatorial search for brevity: exhaustive computational methods deferred to Part 4, later in the course. Here: organize analytically around Figure 12.5’s two primary internal functions: energy storage and vehicle moving
Key architectural question: is energy-storage an input to vehicle-moving (shared form, additive: parallel hybrid), or do they use separate forms (series hybrid: engine drives a generator that supplies the traction motor and/or charges storage; only the electric drive propels the vehicle)?
Third category: combined hybrid: mixes both, either splitting fuel-converter energy across series+parallel simultaneously (power-split hybrids) or switching between only-parallel/only-series modes. This taxonomy (parallel/series/combined) is exactly what the next figure visualizes

Suggested answer: Use the power paths as the diagnostic. In a parallel hybrid, the engine and electric machine can both contribute mechanical power to vehicle propulsion. In a series hybrid, the engine drives a generator and propulsion is delivered through the electric drive; generated electricity can feed the motor and/or charge storage. A combined architecture can provide both kinds of paths or operating modes. A battery being present does not distinguish them. Trace energy conversion, storage, and mechanical connections rather than relying only on labels such as shared form.

Figure 12.8 · Parallel vs. Series Hybrid Drive

Parallel hybrid: the engine and electric motor can both supply propulsion. Series hybrid: the engine drives a generator, which supplies the electric drive and/or charges storage. The electric motor provides propulsion. Source: Crawley, Cameron & Selva (2016), Fig. 12.8.

Makes the parallel-vs-series distinction fully concrete and visual
Parallel path through the summing junction (Σ): fuel and electric converters can both contribute to the energy load. The paths are additive.
Series path from fuel converter to electric converter: the fuel converter does not mechanically drive the load; it supplies electrical power for electric propulsion and/or recharging. Power need not pass through the battery first.
Have students trace both paths visually: solid arrows = propulsion power flow, dashed = recharging flow. This additive/simultaneous vs. sequential/indirect distinction is the basis for the hybrid categories that follow

Discussion: Must a series hybrid send generated electricity through the battery first?

Suggested answer: No. The engine-generator can supply the traction motor and/or charge storage. The distinguishing feature is that the electric drive provides propulsion without a direct engine-to-wheel mechanical path.

Figure 12.9 · The HEV Conceptual Solution Space

Two dimensions organize the space: electric range and degree of electrification (ratio of peak electric motor power to maximum combined power). Seven integrated concepts populate this space: from conventional ICE (0,0) to battery EV (degree = 1). Source: Crawley, Cameron & Selva (2016), Fig. 12.9, after Gorbea (2011).

Payoff of Step 3 fully realized: two-dimensional solution space combining energy-storage/vehicle-moving fragments (parallel/series/combined) with the seven additional fragments (start-stop, regenerative braking, etc.)
Horizontal axis: degree of electrification: ratio of cumulative peak electric-motor power to max combined power. 0 = purely conventional ICE, 1 = purely battery-electric
Vertical axis: electric range: how far on electrical propulsion alone
Regions on the map (detailed next slide): conventional ICE/micro hybrid at the origin; mild hybrid further along electrification, still minimal range; full hybrid moderate electrification, short range; PHEV higher electrification + materially larger range, external charging; series-PHEV (range extender) very high electrification + substantial range, small backup engine; BEV and FCEV both at degree = 1, differing in how electricity is produced (stored battery vs. fuel cell converting on the fly)

Discussion: Does increasing electric range identify the entire hybrid architecture?

Suggested answer: No. Range is one characteristic; power paths, charging arrangements, storage, and control can differ. Distinguish performance attributes from the underlying concept.

Seven Integrated Vehicle Concepts

Concept	Degree of Electrification	Electric Range
Conventional ICE	0	none
Micro Hybrid	minimal	none (start-stop only)
Mild Hybrid	limited	limited or none; model-dependent
Full Hybrid	10–30%	500 m – 3 km
Plug-In Hybrid	>35%	5–160 km
Battery EV (BEV)	1	battery-dependent
Fuel Cell EV (FCEV)	1 (electric propulsion)	not quantified here

Seven integrated concepts populating the solution space, precise numbers attached: these feed into the Pugh matrix a few slides ahead
Conventional ICE: current dominant architecture (diesel/Otto-cycle), electrification and electric range both exactly zero
Micro Hybrid: propulsion exclusively via ICE + motor start-stop, needs a more robust 12-16V battery and starter-generator: “advanced ICE cars,” minimal electrification, no electric-driving capability
Mild Hybrid: some limited electric-driving functionality (though some still don’t offer electric-only driving): high-voltage battery, electric starter-motor for assist, coordinating control system; solely parallel, motor-assist + expanded regenerative braking
Full Hybrid: 10-30% electrification, short electric-only range (~500m-3km): ICE remains primary, electric system materially assists; parallel or combined configs, full regenerative braking
Plug-In Hybrid (PHEV): differentiated by external charging: spans parallel/series/combined, >35% electrification, 5-160 km electric range; series-PHEV range-extender engine can be substantial or a small “limp home” engine
Battery EV (BEV): outside the HEV solution space, electrification = 1, no ICE at all, plug-in by definition, range depends entirely on battery capacity
Fuel Cell EV (FCEV): electrochemical cell converts fuel to current directly; primarily series architecture: fuel-cell system feeds the electrical system, which alone propels
These seven were generated by combining the two primary internal functions’ form-to-function mappings with the seven additional fragments, combinatorially: real coverage of the concept space for Step 4 next

Discussion: Why keep the seven integrated concepts distinct during comparison?

Suggested answer: Their combinations of propulsion and supporting functions differ. A common label such as ‘hybrid’ can hide meaningful differences in operation, infrastructure, and stakeholder value.

Preserve Alternatives and Check Biases

A decision with one option is an imperative: keep several distinct concepts long enough to compare them fairly
Familiar or memorable concepts can feel stronger merely because they are easier to recall
High-fidelity sketches can appear better supported than less-developed ideas, even when the underlying concept is not better
Confirmation and anchoring can make a team favor its first promising candidate

Tip

Use the same goals, reference concept, and evidence standard for every candidate; record uncertain judgments.

Step 4: Select Concepts with a Pugh Matrix

A Pugh matrix rates each candidate concept against a set of criteria: here, the goals from Chapter 11: on a five-level scale from very advantageous (++) to many disadvantages (−−), relative to a chosen reference concept.

Two screening criteria: a backward-looking comparison against the prioritized goals, and a forward-looking judgment of the concept’s potential for a good, elegant architecture
Not every goal differentiates: a goal met equally by every concept (like “must accommodate a driver”) doesn’t help choose among them; it belongs to detailed design more than to the concept decision

Step 4: select a few integrated concepts from the seven generated in Step 3
Why down-selection is necessary: a decision with only one option isn’t a decision. But the architect’s role is to preserve options as long as reasonably possible, resisting pressure to down-select too early before concepts are detailed enough to fully judge fit against goals
Human-judgment subtlety (citing Bazerman): systematic biases toward concepts easier to recall, more information/fidelity, confirming existing beliefs, or already familiar/anchored: none about which concept is actually best. Exactly why a structured tool like Pugh matters: forces criterion-by-criterion comparison, not gut call
Pugh matrix defined: rate each candidate against criteria on a five-level scale (++ to −−), always relative to a chosen reference concept (full HEV serves as reference here, rated “o” against itself by construction)
Two screening considerations: backward (fit against Ch. 11’s prioritized goals: critically/importantly/desirably tiers) and forward (potential for good, elegant architecture)
Important filter: not every goal differentiates. “Must accommodate a driver” is met equally by all seven: constrains detailed design later, doesn’t help choose among these concepts now, so set it aside from this comparison

Discussion: What must be fixed before using a Pugh matrix?

Suggested answer: Define the reference concept, evaluation criteria, and meaning of better, same, or worse. Apply a consistent comparison basis and record uncertain judgments.

Table 12.1 · Critically Important Goals

Goal	ICE	Micro	Mild	Full HEV	PHEV	FCEV	BEV
Environmental satisfaction and impact	−−	−−	−	o	+	++	++
Stable supplier relationships and revenue	o	o	o	o	o	o	o
Accommodate a driver	o	o	o	o	o	o	o

Full parallel HEV is the reference. ++ very advantageous; + some advantages; o average; − some disadvantages; −− many disadvantages. Source: Crawley, Cameron & Selva (2016), Table 12.1.

Table 12.1 · Important Goals

Goal	ICE	Micro	Mild	Full HEV	PHEV	FCEV	BEV
Transport range	+	+	+	o	o	—	−−
Regulatory compliance	o	o	o	o	o	o	o
Carry passengers	o	o	o	o	o	o	o
Stable, rewarding employment	o	o	o	o	o	o	o
Fuel efficiency	−−	−−	−	o	+	++	++

The book leaves FCEV transport range unscored (—); it is not a negative rating. Source: Crawley, Cameron & Selva (2016), Table 12.1.

Table 12.1 · Desirable Goals

Goal	ICE	Micro	Mild	Full HEV	PHEV	FCEV	BEV
Modest investment, volume, return	++	++	+	o	−−	−−	−−
Desirable handling	+	+	o	o	−	−−	−−
Carry cargo	+	+	o	o	−	−−	−
Inexpensive purchase price	++	+	+	o	−	−−	−−

The cargo goal is retained because its scores differ among concepts. Source: Crawley, Cameron & Selva (2016), Table 12.1.

Reading the Pugh Matrix

The benefits of electrification: reduced tank-to-wheel emissions, better fuel consumption, an enhanced ecological image, government incentives
The disadvantages: increased weight, reduced range, higher manufacturing costs, and commercial risk tied to servicing a first-generation high-voltage battery
Beyond the qualitative Pugh scan, one could apply a quantitative comparison for the finalists: covered in Part 4

Tip

A subjective down-selection according to criteria like elegance: is the function-to-form mapping simple and pleasing, closer to one-to-one?: is often the final step in choosing 2–3 concepts for further development.

Draw out the narrative behind the numbers: more memorable tied to real underlying causes
Benefits behind electrified concepts’ good scores: reduced tank-to-wheel emissions, better fuel consumption, enhanced ecological brand image, government financial incentives
Disadvantages behind poor scores in some electrified concepts: increased weight, reduced range for certain storage choices, higher manufacturing costs, and commercial risk servicing a first-generation high-voltage battery. Do not infer an FCEV range score from the blank book cell.
Scope-setting preview: this qualitative Pugh scan is a coarse first pass: once narrowed to finalists, quantitative comparison becomes valuable, covered in detail in Part 4 (Ch. 14-16)
“Elegance”: subjective but useful final screen: is the function-to-form mapping simple and pleasing, closer to one-to-one (Suh’s axiomatic design, Ch. 8) rather than tangled many-to-many? Often the actual final step narrowing to 2-3 concepts in practice: nice bridge into Ch. 13’s decomposition and elegant architecture

Suggested answer: A simple, coherent function-to-form mapping can make a concept easier to explain, integrate, and evolve, but one-to-one allocation is not a sufficient or universal quality test. Shared resources can reduce mass or cost, while extra separation can improve isolation. Use elegance alongside the explicit stakeholder criteria and examine assumptions behind each qualitative score. The matrix is an early comparison relative to a reference, not proof of a quantitatively optimal concept or a timeless ranking of vehicle technologies.

Section 12.3–12.6 Summary

The Hybrid Car concept was expanded via its propulsion fragment (monovalent/bivalent/multivalent) and seven additional internal-function fragments (start-stop, regenerative braking, boosting, load-level increase, electric driving, external charging, gliding)
Fragments were recombined systematically along the parallel/series mapping to produce seven integrated concepts, organized by electric range and degree of electrification
A Pugh matrix, screened backward against prioritized goals and forward against architectural elegance, narrows the field to a small number for further development

Review recapping the fully-worked Hybrid Car example in three sentences, each tied to a framework step
Step 2 (expand/develop fragments): “driving” expanded via propulsion classification (monovalent/bivalent/multivalent) + seven additional fragments (start-stop through gliding)
Step 3 (evolve/refine integrated concepts): fragments recombined systematically via the parallel-vs-series mapping question → seven integrated concepts in the electric-range/electrification solution space
Step 4 (select a few): Pugh matrix, backward (goal-fit) + forward (elegance-potential) screening → small set of finalists
Good template for students to reuse when applying this framework to their own concept-generation exercise later

Discussion: How do the hybrid example’s steps connect?

Suggested answer: Generate propulsion and other fragments, combine them into coherent concepts, then compare the integrated candidates against goals. Each step may reveal a reason to revisit an earlier one.

Figure 12.10 · The Full Funnel

1 solution-neutral function → N concepts (Step 1) → (3–4)N concept fragments (Step 2) → (3–4)N! combinatorial integrated concepts (Step 3) → 10–20 qualitative down-select → 5–7 quantitative down-select → 2–3 final concepts (Step 4). Source: Crawley, Cameron & Selva (2016), Fig. 12.10.

The Figure 12.2 diamond, now with real numbers at every stage: go through each transition slowly, the arithmetic carries the argument
Start: 1 solution-neutral function → Step 1: N distinct concepts → Step 2: each expands into ~3-4 fragments, giving (3-4)×N total fragments → Step 3: combined combinatorially: note the factorial, (3-4)N!: the mathematical reason the space explodes substantially, why Part 4’s computational search methods matter
Step 4 narrows back down in stages: qualitative down-select (Pugh-style) → ~10-20 candidates; quantitative down-select (Part 4 methods) → ~5-7; final outcome → 2-3 concepts for detailed development
Generate alternatives, then narrow the set through successive qualitative and quantitative comparisons.

Discussion: Why should the funnel retain diversity as it narrows?

Suggested answer: Prematurely keeping only similar concepts can discard a family that performs better under revised assumptions. Narrow using evidence while preserving meaningful alternative mechanisms.

Chapter 12 Summary

Concept development combines brainstorming with structured methods such as recombination, TRIZ, and completeness frameworks
The framework expands the alternatives, develops their fragments, combines compatible choices, and selects concepts against the goals
With concept chosen, the architect’s remaining task is managing the resulting investment in complexity: the subject of Chapter 13, where decomposition becomes the primary tool

Tip

Reference: Crawley, E., Cameron, B., & Selva, D. (2016). System Architecture: Strategy and Product Development for Complex Systems. Pearson. Chapter 12.

Close main content, tying back to Figure 12.1’s three themes before the closing case study
Brainstorming and structured methods can contribute to the same concept-development process.
Review how the four steps connect concept generation to selection in Figure 12.10.
Looking forward: once concept is chosen, the architect’s task shifts to managing the resulting complexity investment: Ch. 13’s subject, decomposition as the primary tool, continuing Figure 12.1’s downward arrow from concept toward architecture/operations

Discussion: What is a useful stopping point for this chapter’s concept work?

Suggested answer: A small set of coherent candidates with documented assumptions and reasons for retention. Detailed quantitative trade studies can then focus effort on the most promising alternatives.

Box 12.2 · Case Study: Architectural Competition in the Automotive Industry

Early Architectural Competition

Architectural competition refers to differentiating a product from others in the market based on product architecture, rather than incremental improvement within a single dominant one.

In the automotive industry’s early years, three different concepts: electric, steam, and internal combustion: competed to dominate the market
Electric cars were marketed to female drivers for ease of use and minimal maintenance; ICE cars targeted male drivers seeking power and speed; steam cars promised long range on a single “filling of the tanks”

Introduce the closing case study: credit Dr. Carlos Gorbea’s contribution (book acknowledges his dissertation on vehicle architecture and lifecycle cost analysis): model good academic citation practice
Architectural competition, defined: firms differentiating products by different architecture, not incremental improvement within one dominant architecture: contrasted with the later “architectural dominance” period (subsystem-level competition)
Historical stage: automotive industry’s early years: electric, steam, and internal combustion competed for dominance, each linking powertrain elements differently to achieve the same core function
Marketing segmentation (memorable, concrete): electric marketed to female drivers (ease of use, minimal maintenance); ICE targeted male drivers (power, speed); steam promised long range on one “filling of the tanks” (note: plural: both fuel and water tank, relevant again shortly)

Discussion: Why can competing architectures coexist early in an industry?

Suggested answer: Different concepts satisfy different use cases while technology, infrastructure, and preferences remain unsettled. Early success in one context does not establish universal dominance.

Figure 12.11 · Marketing an Early Electric Car

The Baker Electric Runabout marketed as “The Aristocrats of Motordom”: evidence of genuine architectural competition in the early automotive market, each architecture appealing to a different customer segment. Source: Crawley, Cameron & Selva (2016), Fig. 12.11. Source: Baker Electrics.

Primary-source evidence for the previous slide’s claims, not decoration
Tagline: “The Aristocrats of Motordom”: pitch: constant speed/mileage increases, reliability no longer limited by charging-station availability, “the ideal car for town and suburban service”: refined, low-maintenance, genteel customer base
Contrast: a 1904 steam-car ad (referenced in the same passage) pitched in technical language: 100 miles of range on one filling of “the tanks” (plural: fuel + water). Same aggregate function (driving), different architectures, visibly different customer sensibilities: clean illustration of architectural competition in an actual marketplace

Discussion: What can an early electric-car advertisement tell us about architecture?

Suggested answer: It reveals benefits the seller chose to emphasize and assumptions about users. It does not independently establish comparative performance or explain the entire market outcome.

What Triggered ICE Dominance?

Early steam cars had rapid acceleration but needed frequent water refills and 20 minutes to build boiler pressure; early electrics were simple but limited to ~64 km range and 32 km/h; ICE cars matched electrics in performance but were harder and riskier to start (hand crank)
Two breakthroughs resolved steam’s core weakness (water dependency): the internal combustion engine and the electric motor: both first proven in rail and power generation before entering automobiles
Ford’s assembly line (lower price) and the electric starter (removed the crank-injury risk) made ICE cars affordable and safe for the masses by 1920: steam disappeared from the market entirely by 1930

Three-way performance comparison: explains why the outcome wasn’t obvious in advance
Steam: best raw acceleration (0-65 km/h in under 10s, peak torque from a standing start), but frequent water refills (~2 miles between fills on early models) and up to 20 minutes to build boiler pressure
Electric: comparable price to steam, needed a power source mostly available in major cities; range <64 km, top speed ~32 km/h (low energy density: ~15-20 Wh/kg lead-acid vs. gasoline’s ~12,200 Wh/kg): but simple to operate, no complicated shifting, no maintenance/emissions
ICE: performance on par with electric, inferior to steam: steam proponents called it the “internal explosion engine” to imply danger, with real basis: startable in under a couple minutes, but many motorists were injured hand-cranking it
Two breakthroughs resolved steam’s core weakness (water dependency): internal combustion engine and electric motor: both first proven in rail and stationary power generation before entering automobiles
What actually tipped it to ICE: Ford’s assembly line (lower price) + electric starter (removed hand-crank injury risk) → affordable and safe by ~1920. Steam, despite a lineage back to the 1780s, disappeared from the market entirely by 1930

Discussion: Why avoid explaining ICE dominance with one component advantage?

Suggested answer: Adoption depends on interacting performance, cost, infrastructure, production, and user factors. A system-level explanation should connect these conditions rather than treating one invention as sufficient.

Figure 12.12 · A Century of Architectural Competition

Three eras: initial architectural competition (1885–1915), dominance of the ICE architecture (1915–1998), punctuated by regulatory pressure and GM’s EV-1, and renewed architectural competition (1998–2008) beginning with the Toyota Prius and modern EVs. Source: Crawley, Cameron & Selva (2016), Fig. 12.12, after Gorbea, Fricke & Lindemann (2008).

Visual centerpiece tying the whole case study together: three named eras
Initial architectural competition (1885-~1915): steam, electric, ICE compete as different architectures
Dominant-architecture era: ICE design principles relatively unchanged since the mid-1930s. Once the market adopted one architecture, the risk of not knowing the winner disappeared: manufacturers redirected innovation to the subsystem level, exactly the “incremental improvement within a dominant architecture” contrast from the case study’s opening slide
- Triggering events annotated: Ford’s assembly line (price); steam’s failed 1920s-1930 last stand (water-recycling condenser, faster-pressure spark-ignition starter: made it heavier/more expensive, gone by 1930); rising regulatory requirements (safety, emissions); GM’s EV-1 as an early, ultimately unsuccessful disruption attempt
Renewed architectural competition (~1998-2008): reintroduction of EVs + Toyota Prius (first mass-produced HEV in the U.S.): book frames this as a return to genuine architectural, not only incremental, competition, for the first time in nearly a century

Discussion: What changes when competition moves from architecture to components?

Suggested answer: Firms refine subsystems within a broadly shared configuration. A later shift in requirements or technology can reopen the architectural choices and challenge established capabilities.

The Lesson of Dominant Architecture

Once the market adopted ICE as the single dominant architecture, the risk of not knowing which architecture would prevail was eliminated: manufacturers could focus innovation at the subsystem level, not the architecture level
Decades of incremental innovation followed: most automakers built core competencies deeply specific to ICE design, leaving them poorly positioned to pivot when architecture became contested again

Renewed architectural competition can challenge firms whose skills and investments depend on the established design. The decline of steam-car manufacturers illustrates the need to reassess those assumptions.

The case study’s result slide: strategic business lesson
Once the market settled on ICE, the risk of not knowing the winning architecture was eliminated for everyone simultaneously: industry redirected innovation to the subsystem level rather than the riskier question of overall architecture
The real strategic warning: decades of incremental innovation meant automakers built core competencies: expertise, supplier relationships, manufacturing investment: deeply around ICE design, leaving them structurally poorly positioned to pivot when architectural competition returned
Ask students how an established manufacturer’s expertise and investment could help or hinder a change in architecture.

Suggested answer: Yes in principle: an architecture that once lost can become competitive when component capabilities, infrastructure, user needs, or lifecycle economics change. Evaluate electric propulsion against the relevant present-day use case and alternatives rather than assuming that the historical winner must remain dominant. This question illustrates conditional architectural competition, not a forecast of one winner in every market. Incumbent expertise can be valuable yet also make a transition difficult; new competition does not imply that all previous competencies become useless.

Next: Decomposition as a Tool for Managing Complexity

Chapter 13 examines how decomposition helps manage the complexity of developing a selected concept.