2.1 Technical Feasibility Studies & Automation Architecture Alternatives

Engineering Feasibility and Control System Architecture Selection

In the initial lifecycle phases governed by the ISA Certified Automation Professional body of knowledge (Domain 1: Feasibility Study), selecting an automation platform is among the most consequential technical decisions an engineer faces. An inappropriate architecture selected during conceptual screening introduces severe operational bottlenecks, unmanageable lifecycle maintenance costs, or outright project failure. The objective of a technical feasibility study is not simply to identify hardware that can physically interface with field instrumentation, but to evaluate how well architectural alternatives satisfy the fundamental operating characteristics of the target process.

Automation systems are broadly grouped into four architectural families:

1. Distributed Control Systems (DCS): Engineered natively for large-scale, continuous process facilities (refineries, petrochemicals, large power utilities) where regulatory analog control loops, unified tag databases, integrated alarm management, and high availability take precedence.
2. Programmable Logic Controllers (PLC) and Programmable Automation Controllers (PAC): Engineered natively for high-speed discrete manufacturing, assembly, packaging, and unit machine control requiring rapid scan execution (typically $1\text{ to }10\text{ ms}$), modular I/O, and flexible programming under IEC 61131-3 standards.
3. Supervisory Control and Data Acquisition (SCADA): Engineered for wide-area, geographically distributed infrastructure (pipelines, municipal water distribution, electrical transmission grids) where communication links may be intermittent, high-latency, or low-bandwidth, and remote assets require autonomous Remote Terminal Units (RTUs).
4. Industrial PCs (IPC) and Embedded Soft-PLCs: Engineered for computationally intensive specialized applications, including machine vision inspection, edge AI/ML inference, high-axis coordinated motion, and advanced algorithmic processing running on hard real-time operating system (RTOS) kernels.

Platform Comparison Matrix

---

Architectural Trade-Off Analysis

Determinism, Scan Cycles, and Execution Timing

A critical distinction on the ISA CAP exam is the difference between deterministic execution and best-effort processing. A control loop is deterministic if its execution time is mathematically bounded and guaranteed within a fixed schedule.

• Continuous Processes: Thermal inertia and fluid hydraulic dynamics typically exhibit process dead times ($t_d$) and time constants ($\tau$) measured in seconds or minutes. For these systems, a controller loop execution period of $100\text{ to }250\text{ ms}$ provides adequate controllability without overloading controller processors.
• High-Speed Discrete Processes: Packaging lines, parts sorting, and mechanical stamping require sensor evaluation and solenoid actuation within $2\text{ to }10\text{ ms}$. If a controller exhibits scan-time jitter of $\pm 15\text{ ms}$, parts miss registration, causing mechanical jams.
• Operating System Architecture: PLCs and DCS controllers execute proprietary firmware running on deterministic RTOS kernels (such as VxWorks or QNX). Industrial PCs hosting "soft-logic" engines require dedicated real-time extensions (e.g., hypervisor partitioning) that isolate control loops from non-deterministic general-purpose operating system (GPOS) tasks such as Windows background services or file system logging.

Database Tag Management and Engineering Overhead

In a DCS, the controller code, I/O assignment, engineering units, alarm trip points (ISA-18.2), and HMI display attributes reside in a single global object-oriented database. Creating an analog input block automatically instantiates the tag, assigns faceplate graphics, establishes trending, and configures historical data collection. In large plants with thousands of I/O points, this unified model substantially reduces configuration and tag-maintenance effort.

In contrast, traditional PLC/PAC architectures separate the controller tag engine from the supervisory HMI software. Modifying a high-temperature alarm limit in the controller requires manual or scripted synchronization with the HMI tag dictionary. For small systems ($<1{,}000\text{ I/O}$), this separation offers flexibility and lower software licensing costs. However, in large continuous processes, maintaining dual databases introduces substantial configuration management risk and human error.

Redundancy Architectures and Failure Modes

System availability directly dictates plant uptime. Technical feasibility assessments must quantify the mean time between failures (MTBF) and mean time to repair (MTTR) for control architectures:

• 1:1 Hot Standby Redundancy: Dual controllers execute code synchronously. The secondary processor continuously mirrors internal state variables, timers, and I/O image tables across a dedicated high-speed fiber-optic synchronization link. If the primary processor fails (watchdog timeout, hardware fault, power loss), the secondary assumes control bumplessly within $10\text{ to }20\text{ ms}$ without upsetting field actuators.
• Warm Standby: The backup controller is powered on but does not mirror dynamic memory state. Upon primary failure, the backup initializes its executive program, reads current field I/O states, and begins execution after a delay of $500\text{ ms to }5\text{ s}$. This latency is unacceptable for continuous reactors or critical compressors, but acceptable for slow thermal furnaces or wastewater aerators.
• Triple Modular Redundancy (TMR): Employs three isolated processing channels executing identical logic in parallel with two-out-of-three (2oo3) hardware voting. TMR architectures tolerate a single hardware fault without losing operational continuity and are standard in Safety Instrumented Systems (SIS) and turbomachinery governors.

---

Plant Physical Infrastructure Constraints

A technically viable automation architecture cannot function without adequate physical and environmental support. During the feasibility study, automation engineers must verify physical space, power supply stability, thermal management, environmental air quality, and cabling infrastructure.

Loading diagram...

Heat Dissipation and Thermal Loading Calculations

Industrial control cabinets house heat-generating hardware: switching power supplies, processor modules, I/O field terminal blocks, network switches, and intrinsic safety isolation barriers. Standard electrical efficiency dictates that virtually all electrical power consumed by control equipment inside an enclosure converts to thermal energy.

To size control room HVAC cooling capacity, calculate the total heat dissipation using the thermal conversion equation:

$Q_{\text{BTU/hr}} = 3.41214 \times P_{\text{watts}}$

Where:

• $Q_{\text{BTU/hr}}$ is the sensible heat dissipation rate in British Thermal Units per hour
• $P_{\text{watts}}$ is the total active electrical power consumed inside the enclosure in watts

Good practice adds a future growth margin (commonly 20% to 25%, set by the project specification) to the calculated equipment power load before final HVAC specification. Failure to manage internal cabinet temperatures shortens semiconductor life; a common Arrhenius-based rule of thumb is that electronic failure rates roughly double for every $10^\circ\text{C}$ rise in operating temperature.

Clean Instrument Power and Grounding Architectures

Automation systems require high-integrity "clean" power isolated from inductive electrical loads (large electric motors, variable frequency drives, arc welders) that induce electrical noise, harmonics, and voltage sags:

• Online Double-Conversion UPS: Feasibility studies must specify true online double-conversion UPS systems. The utility AC supply is continuously converted to DC to charge batteries and supply an inverter that generates a clean, regulated sinusoidal AC output. There is zero transfer time ($0\text{ ms}$) during a utility power failure, preventing controller power-supply dropouts.
• UPS Sizing Formula: Apparent power rating ($S_{\text{UPS}}$) must account for the equipment power factor ($\text{PF}$) and future growth margin ($M_{\text{growth}}$):

$S_{\text{UPS}} = \frac{P_{\text{load}}}{\text{PF}} \times (1 + M_{\text{growth}})$

• Signal Grounding vs. Safety Grounding: Control systems mandate a Single-Point Ground (SPG) topology—frequently called an Isolated Instrument Ground (IIG). Instrument cable shields must be grounded at one end only (typically at the control cabinet marshalling terminal strip) to prevent ground loops caused by earth potential differences between field transmitters and the control room.

Environmental Airborne Contaminants (ANSI/ISA-71.04)

In chemical, pulp and paper, and refining plants, atmospheric corrosive gases such as hydrogen sulfide ($\text{H}_2\text{S}$), sulfur dioxide ($\text{SO}_2$ circuit-board corrosion), and chlorine ($\text{Cl}_2$) react with copper and silver electrical traces. ANSI/ISA-71.04 classifies environmental severity into four standard categories using reactive coupon monitoring:

1. Class G1 (Mild): Copper reactivity $<300\text{ \AA/30 days}$; silver reactivity $<200\text{ \AA/30 days}$. Well-controlled office or high-grade control room. Unprotected commercial electronics survive.
2. Class G2 (Moderate): Copper reactivity $300\text{ to }1{,}000\text{ \AA/30 days}$; silver reactivity $200\text{ to }1{,}000\text{ \AA/30 days}$. Light industrial environment; copper creep corrosion begins on exposed circuit traces.
3. Class G3 (Harsh): Copper reactivity $1{,}000\text{ to }2{,}000\text{ \AA/30 days}$; silver reactivity $1{,}000\text{ to }2{,}000\text{ \AA/30 days}$. High probability of corrosive failure in standard electronics; protection such as gas-phase filtration of the room air and conformally coated electronics is normally needed.
4. Class GX (Severe): Copper reactivity $\ge 2{,}000\text{ \AA/30 days}$; silver reactivity $\ge 2{,}000\text{ \AA/30 days}$. Electronics will fail rapidly. Control equipment needs a pressurized room or enclosure supplied with chemically filtered air (for example, potassium permanganate and activated carbon media).

Industrial Transmission Media: Copper vs. Fiber Optics

When evaluating field network topology during feasibility, the physical distance between field marshalling cabinets and central control rooms determines the transmission medium:

• Category 6A Shielded Twisted Pair (STP): Maximum transmission distance is strictly limited to $100\text{ m}$ ($328\text{ ft}$) for industrial Ethernet (100BASE-TX / 1000BASE-T). Copper cable is susceptible to electromagnetic interference (EMI) and radio frequency interference (RFI) when routed near high-voltage motor feeder cables.
• Multi-Mode Fiber (MMF — OM3 / OM4): Utilizes a larger core ($50/125\text{ }\mu\text{m}$) driven by $850\text{ nm}$ vertical-cavity surface-emitting lasers (VCSELs). Maximum distance for gigabit transmission is $300\text{ to }550\text{ m}$. Ideal for intra-building or localized process unit backbones.
• Single-Mode Fiber (SMF — OS2): Utilizes an ultra-thin core ($9/125\text{ }\mu\text{m}$) driven by $1310\text{ nm}$ or $1550\text{ nm}$ solid-state laser diodes. Transmits optical signals up to $10\text{ to }40\text{ km}$ without repeaters. Fiber optics are 100% dielectric, providing total immunity to EMI, RFI, and ground potential differences.

The optical link loss budget equation verifies link feasibility before detailed design:

$\alpha_{\text{total}} = (L \times \alpha_{\text{cable}}) + (N_{\text{splices}} \times \alpha_{\text{splice}}) + (N_{\text{connectors}} \times \alpha_{\text{conn}}) + M_{\text{safety}}$

Where $\alpha$ represents optical attenuation in decibels ($\text{dB}$), $L$ is fiber distance in kilometers, and $M_{\text{safety}}$ is an engineering design margin (typically $3\text{ dB}$).