1.2 System Analysis Tools, Process Definition, Requirements Engineering & QFD

System Analysis Tools, Process Definition, Requirements Engineering & QFD

Core Principle: Industrial and systems engineers optimize systems by first rigorously defining operational processes, decomposing stakeholder requirements, and systematically mapping qualitative customer desires into measurable, quantifiable engineering parameters through Quality Function Deployment (QFD).

Before an industrial system can be optimized, simulated, or automated, its architecture must be precisely documented. Process definition establishes system boundaries, interfaces, inputs, and outputs. Requirements engineering ensures that engineering specifications directly reflect true user demands rather than assumed technical solutions. The central analytical framework connecting customer requirements to engineering design is the House of Quality (HOQ).

---

1. Process Mapping Frameworks

Different engineering questions require different modeling tools. The four primary process mapping methodologies encountered in industrial systems design are contrasted below:

SIPOC Analysis

A SIPOC (Suppliers, Inputs, Process, Outputs, Customers) diagram is constructed during initial project definition to prevent scope creep. It identifies:

• Suppliers: Entities providing raw materials, information, or subassemblies.
• Inputs: Physical materials, data, specifications, energy, or labor.
• Process: A macro-level summary consisting of 4 to 7 core transformation steps.
• Outputs: The resulting physical products, services, documentation, or emissions.
• Customers: The internal downstream workstations or external purchasers receiving the outputs.

Value Stream Mapping and Process Cycle Efficiency

A Value Stream Map (VSM) captures both material flows and information flows across an entire facility. The lower portion of a VSM features a timeline ladder:

• The upper notches of the ladder represent queue times, buffer inventory dwell times, and transit delays (Non-Value-Added Time, $NVA$).
• The lower troughs represent actual machine or operator transformation times (Value-Added Time, $VA$, or processing time $PT$).

The fundamental metric derived from a VSM is Process Cycle Efficiency ($PCE$), also termed Process Velocity:

$PCE = \frac{\sum \text{Value-Added Processing Time}}{\text{Total Production Lead Time}} \times 100\%$

Where: $\text{Total Production Lead Time} = \sum \text{Value-Added Time} + \sum \text{Non-Value-Added Delay Time}$

In typical un-lean industrial environments, $PCE$ is frequently below 5%, revealing that workpieces spend 95% or more of their total factory lifespan waiting in queues, batch staging, or transport buffers.

---

2. Requirements Engineering and the Kano Model

Requirements engineering translates qualitative stakeholder desires into verifiable engineering specifications. The Voice of the Customer (VOC) is captured through ethnographic observation, customer interviews, warranty claim analytics, and focus groups.

The Kano Model of Customer Satisfaction

Dr. Noriaki Kano classified customer requirements into distinct categories based on their relationship to customer satisfaction:

Customer Satisfaction vs Attribute Execution (Kano Model)
                 Satisfied ^
                           |          /  Attractive (Delighters)
                           |         /   [Unexpected innovations]
                           |        / 
                           |       /   / One-Dimensional (Performance)
                           |      /   /    [Higher execution = higher satisfaction]
                           |     /   /
    Absent                 |    /   /                     Fully Implemented
    <----------------------+--------------------------------------------->
                           |  /   / 
                           | /   /
                           |/___/___ Must-Be (Basic / Threshold)
                           |        [Expected baseline; cause severe dissatisfaction if absent]
               Dissatisfied v

1. Must-Be (Threshold / Basic) Requirements: Non-negotiable baseline characteristics. If absent or failing, the customer is profoundly dissatisfied. However, executing them beyond standard expectations does not increase satisfaction (e.g., a commercial aircraft door sealing properly, or an automobile braking system functioning).
2. One-Dimensional (Performance) Requirements: Satisfaction is linearly proportional to the level of execution. Better performance yields higher satisfaction (e.g., battery operating life, fuel economy, processing speed).
3. Attractive (Excitement / Delighters) Requirements: Unexpected innovations that generate disproportionate delight when present, but cause zero dissatisfaction when omitted because customers did not anticipate them (e.g., self-cleaning nozzle technology, automated predictive maintenance alerts).
4. Indifferent Attributes: Features for which customers have no preference; investing engineering capital here yields no commercial return.

Axiomatic Design Principles

In systems requirements engineering, Nam Pyo Suh's Axiomatic Design establishes two core design axioms:

• Axiom 1: The Independence Axiom: Maintain the independence of functional requirements ($FRs$). In an acceptable design, the mapping between functional requirements ($FRs$) and physical design parameters ($DPs$) must be decoupled or uncoupled, such that modifying a design parameter satisfies one functional requirement without perturbing another.
• Axiom 2: The Information Axiom: Minimize the information content of the design. Among designs that satisfy the Independence Axiom, the optimal design possesses the minimum information content (i.e., highest probability of successful manufacturing and operational reliability).

---

3. Quality Function Deployment (QFD) & The House of Quality

Quality Function Deployment (QFD) is a structured systems methodology developed by Yoji Akao in 1966. Its primary analytical matrix is the House of Quality (HOQ), which systematically translates customer requirements into technical product specifications.

Anatomy of the House of Quality

                      / \       Roof: Correlation Matrix
                     /   \      [Technical Trade-offs: ++, +, -, --]
                    /_____\
                   |       | 
    +--------------+-------+--------------+
    | Customer     |  (3)  |     (5)      |
    | Requirements |  REL  |  Competitive |
    | (WHATs)      | MATRIX|  Assessment  |
    |     &        |       |  (Benchmarking)|
    | Importance   | 9-3-1 |     1 to 5   |
    |    (1-5)     |       |              |
    +--------------+-------+--------------+
                   |  (4)  |
                   | TECH  |
                   | SPECS |
                   | (HOWs)|
                   +-------+
                   |  (6)  |
                   | SCORES|
                   | W_j   |
                   +-------+

The HOQ matrix is organized into six functional zones:

1. Customer Requirements (WHATs): Structured list of customer needs categorized into primary, secondary, and tertiary levels, paired with customer importance weights ($I_i$), typically rated on a scale of 1 (low importance) to 5 (critical importance).
2. Technical Descriptors (HOWs): Engineering parameters and design characteristics that directly influence one or more WHATs. Each HOW must be objectively measurable and carry an optimization vector:
   • Maximize ($\uparrow$)
   • Minimize ($\downarrow$)
   • Target Value ($\odot$)
3. Relationship Matrix: The central grid mapping how strongly each technical characteristic impacts each customer requirement. Conventional NCEES scoring utilizes non-linear weights:
   • Strong Relationship: Symbol $\bullet$, numerical value $= 9$
   • Moderate Relationship: Symbol $\circ$, numerical value $= 3$
   • Weak Relationship: Symbol $\triangle$, numerical value $= 1$
   • No Relationship: Blank cell, numerical value $= 0$

1. Correlation Roof Matrix: The triangular roof mapping interactions between technical characteristics (HOW vs HOW). It documents technical synergies and physical trade-offs:
   • Strong Positive ($++$) or Positive ($+$): Mutual synergy (e.g., increasing material strength also increases wear resistance).
   • Negative ($-$) or Strong Negative ($--$): Engineering conflict (e.g., increasing structural wall thickness increases part weight and cooling cycle time).
2. Competitive Assessment: Benchmarking the current product against competitor offerings across both customer perceptions (1–5 scale) and technical performance metrics.
3. Technical Importance Scores and Priorities:
   • Absolute Technical Weight ($W_j$): The weighted sum of relationship values for technical characteristic $j$ across all $m$ customer requirements:

$W_j = \sum_{i=1}^m \left( I_i \times R_{ij} \right)$

Where:

• $I_i$ is the customer importance rating of requirement $i$.

• $R_{ij}$ is the numerical relationship score between requirement $i$ and technical characteristic $j$ ($R_{ij} \in \{0, 1, 3, 9\}$).

• Relative Technical Importance ($w_j$): The normalized percentage weight of characteristic $j$ relative to the sum of all $n$ technical characteristics:

$w_j = \frac{W_j}{\sum_{k=1}^n W_k} \times 100\%$

---

4. Comprehensive Worked Numerical HOQ Example

An industrial engineering team is designing a next-generation semi-automated medical packaging station. Through voice of customer interviews, three critical requirements are identified:

• $CR_1$: Easy to sanitize and clean (Importance $I_1 = 5$)
• $CR_2$: Long operational service life (Importance $I_2 = 4$)
• $CR_3$: Rapid machine cycle time (Importance $I_3 = 3$)

The systems engineering team establishes three measurable engineering parameters:

• $HOW_1$: Surface Roughness $R_a$ (Target: Minimize $\downarrow$, measured in $\mu\text{m}$)
• $HOW_2$: Tensile Yield Strength $\sigma_y$ (Target: Maximize $\uparrow$, measured in $\text{MPa}$)
• $HOW_3$: Actuator Slew Rate $v_{\max}$ (Target: Maximize $\uparrow$, measured in $\text{mm/s}$)

The cross-functional team establishes the following relationship assignments using the 9-3-1 scale:

• $CR_1$ has a Strong relationship with Surface Roughness ($R_{11} = 9$), No relationship with Tensile Strength ($R_{12} = 0$), and a Weak relationship with Actuator Slew Rate ($R_{13} = 1$).
• $CR_2$ has a Moderate relationship with Surface Roughness ($R_{21} = 3$), a Strong relationship with Tensile Strength ($R_{22} = 9$), and a Moderate relationship with Actuator Slew Rate ($R_{23} = 3$).
• $CR_3$ has No relationship with Surface Roughness ($R_{31} = 0$), a Moderate relationship with Tensile Strength ($R_{32} = 3$), and a Strong relationship with Actuator Slew Rate ($R_{33} = 9$).

Matrix Representation

Step-by-Step Mathematical Calculation

Step 1: Compute Absolute Technical Weight for $HOW_1$ (Surface Roughness): $W_1 = (I_1 \times R_{11}) + (I_2 \times R_{21}) + (I_3 \times R_{31})$ $W_1 = (5 \times 9) + (4 \times 3) + (3 \times 0) = 45 + 12 + 0 = 57$

Step 2: Compute Absolute Technical Weight for $HOW_2$ (Tensile Strength): $W_2 = (I_1 \times R_{12}) + (I_2 \times R_{22}) + (I_3 \times R_{32})$ $W_2 = (5 \times 0) + (4 \times 9) + (3 \times 3) = 0 + 36 + 9 = 45$

Step 3: Compute Absolute Technical Weight for $HOW_3$ (Actuator Slew Rate): $W_3 = (I_1 \times R_{13}) + (I_2 \times R_{23}) + (I_3 \times R_{33})$ $W_3 = (5 \times 1) + (4 \times 3) + (3 \times 9) = 5 + 12 + 27 = 44$

Step 4: Compute Total Technical Weight Sum: $\sum_{k=1}^3 W_k = W_1 + W_2 + W_3 = 57 + 45 + 44 = 146$

Step 5: Compute Percentage Relative Importance ($w_j$): $w_1 = \frac{57}{146} \times 100\% = 39.04\%$ $w_2 = \frac{45}{146} \times 100\% = 30.82\%$ $w_3 = \frac{44}{146} \times 100\% = 30.14\%$ $\sum w_j = 39.04\% + 30.82\% + 30.14\% = 100.00\%$

Engineering Interpretation: Surface roughness ($R_a$) accounts for 39.04% of total technical priority, making it the single most critical engineering specification. Machining and finishing processes must prioritize surface polish to satisfy sanitary standards before investing capital in ultra-high yield alloys or high-speed linear servos.

---

5. Common Exam Pitfalls in Requirements & QFD

1. Conflating Customer Needs (WHATs) with Engineering Solutions (HOWs): On exam questions asking candidates to structure an HOQ, placing a technical mechanism (e.g., "Use brushless DC motor") in the customer requirement row is incorrect. The customer need is "Quiet machine operation"; the motor technology is an engineering descriptor (HOW).
2. Applying Linear Weights when 9-3-1 is Specified: If an exam problem explicitly specifies conventional QFD weights, using a linear 3-2-1 scale yields incorrect numerical results. Check whether the problem defines 9-3-1 or 5-3-1.
3. Ignoring Negative Roof Interactions: A technical characteristic with high relative importance may exhibit severe negative correlations with other critical parameters in the roof matrix. Optimizing one parameter without addressing negative roof trade-offs causes systemic failure.
4. Confusing Process Lead Time with Cycle Time in VSM: Total lead time includes the duration workpieces spend sitting in inter-stage inventory buffers. Cycle time refers strictly to transformation time at a processing station.

---

6. System Analysis and Design Tools: Flowcharts, Pareto Charts, Affinity Diagrams, NGT & Input–Output Analysis

The NCEES specification lists these tools by name under systems analysis and design. Each answers a different question, and exam items often ask which tool fits a situation.

Flowcharts

A flowchart uses standard symbols: an oval for start and end, a rectangle for a process step, a diamond for a decision, a parallelogram for input or output, and a document symbol for paperwork. Drawing the current state often reveals rework loops, duplicate approvals, and decisions with no owner. Swimlanes add accountability by placing each step in the lane of the department that performs it.

Pareto charts

A Pareto chart sorts categories (defect types, downtime causes, complaint codes) from largest to smallest and adds a cumulative-percentage line. The usual finding is that a "vital few" categories account for most of the total, which tells the team where to start. Rank by cost or impact, not just count, when categories differ in severity.

Affinity diagrams

The affinity diagram (the KJ method, after Jiro Kawakita) organizes a large set of ideas or customer statements:

1. Write each idea or comment on its own card or note.
2. Team members sort the cards into groups silently, moving cards until the groups are stable.
3. Write a header card that names the theme of each group.
4. Discuss the groups and, if needed, arrange them into a hierarchy.

Affinity diagrams are useful early in requirements work. The grouped voice-of-the-customer statements become the customer requirements (WHATs) of a House of Quality. The specification also lists the affinity diagram under layout design, where it groups activities that belong together before a relationship chart is drawn.

Nominal Group Technique (NGT)

NGT produces a ranked decision while preventing one or two people from dominating:

1. Silent generation: each member writes ideas independently.
2. Round-robin recording: each member states one idea per turn until all are listed, with no debate.
3. Clarification: the group discusses meaning only, not merit.
4. Independent voting: each member ranks a set number of ideas, for example 5 points for the top choice down to 1 point.

Example: Six team members rank five ideas for reducing changeover time. The point totals are A = 22, B = 9, C = 25, D = 18, and E = 16. Idea C is the group's priority even if its sponsor spoke the least during discussion.

Input–Output Analysis

At the system level, input–output analysis defines the system boundary and lists what crosses it: inputs (materials, information, energy), outputs (products, data, waste), controls (standards, schedules), and mechanisms (people, equipment). This is the basis of IDEF0 function models and of the SIPOC diagram above.

At the work-center level, input–output control compares planned and actual hours of work released to and completed by a work center:

$\text{Backlog}_t = \text{Backlog}_{t-1} + \text{Actual Input}_t - \text{Actual Output}_t$

Input is close to plan, but output fell 20 hours behind by week 2, so the backlog and queue time grew. The fix is on the capacity side, such as overtime or reducing downtime at that work center. Releasing more work would only lengthen the queue.