🏭 Generator System Design: Paralleling, Transfer Switches, and Fuel Systems

Sizing the Generator vs. Designing the System Around It

Our Generator Sizing Guide covers how to calculate the kW/kVA a generator needs to deliver — running watts vs. starting watts, NEC Article 220 demand load calculations, and ISO 8528 power rating classes. This article assumes that sizing work is already done and addresses a different question: once you know the generator's kW rating, how do you design the complete system around it — the transfer switching, redundancy strategy, fuel supply, code classification, and load management — that makes a properly sized generator into a reliable backup power system?

Automatic Transfer Switch Types and Transition Modes

The automatic transfer switch (ATS) is the device that senses utility failure, signals the generator to start, and transfers the building's load from utility to generator source (and back, once utility power is restored and stable). Beyond the basic manual-vs-automatic distinction covered in the sizing guide, system design must select a transition mode:

• Open-transition switching — the load is briefly disconnected from both sources during the transfer (a true break-before-make sequence). This is the simplest and most common configuration and is acceptable for the vast majority of emergency and standby applications, but it causes a momentary interruption — typically well under a second, but nonzero — that can disrupt sensitive electronic equipment, cause motor loads to require a re-start inrush, or interrupt processes that cannot tolerate any break in power.
• Closed-transition switching — the ATS briefly parallels the utility and generator sources for a fraction of a second (make-before-break) so the transfer is seamless, with zero interruption to the load. This requires the generator to already be running, synchronized, and stable before transfer, which means closed-transition is used for planned transfers (such as periodic generator testing under load, or scheduled utility maintenance) rather than for the initial response to an unplanned utility outage, since the generator cannot be pre-synchronized to a utility source that has already failed. Closed-transition equipment is more complex and costly than open-transition and requires utility coordination/approval since it briefly parallels with the utility grid.

Facility design should specify open-transition ATS for standard emergency response to utility failure, and add closed-transition capability specifically where seamless planned testing or scheduled transfers are a design requirement — not as a default for every application.

Generator Paralleling for Redundancy and Capacity

Rather than relying on a single large generator, many facilities — hospitals, data centers, and other high-reliability sites — parallel multiple smaller generators onto a common bus. Paralleling serves two distinct design goals simultaneously: redundancy (if one generator in the paralleled set fails to start or trips offline, the remaining units can still carry some or all of the critical load, an outcome a single generator can never provide) and capacity (multiple units can be added incrementally as load grows, or can jointly supply a load larger than any single available generator frame size).

Paralleling generators requires synchronization — each generator's output must match the others in voltage magnitude, frequency, and phase angle before its breaker closes onto the shared bus. Closing a generator breaker out of synchronization produces a severe transient current and mechanical shock to the generator shaft and can cause serious equipment damage. This is the same underlying physics that governs available fault current and interrupting rating coordination elsewhere in a power system — both are ultimately about controlling how much energy is released when electrical connections are made or broken under imperfect conditions, and both require the protective/switching equipment to be rated for the worst-case transient it may see. Modern paralleling switchgear uses automatic synchronizing controls that monitor the incoming generator's frequency and phase relative to the bus and only permit breaker closure once they are matched within a tight tolerance, with a check-synchronizing relay as a hardwired safety backstop.

Fuel System Design

Fuel system design must match the generator's expected run-time requirement to an adequate, appropriately configured fuel supply:

• Diesel day tank vs. bulk storage. A day tank is a smaller sub-tank, local to the generator, that supplies the engine directly and is automatically refilled from a larger bulk storage tank (often underground or a larger above-ground tank) via a transfer pump system. Day tanks provide a buffer that keeps the generator running even briefly if the transfer pump or bulk tank has an issue, and they simplify piping and fuel-quality management at the engine. Bulk storage sizing is driven by the facility's required run-time — the longer the mandated or desired autonomous run-time, the larger the bulk tank, with the day tank sized only for short-term buffering rather than full run-time capacity.
• Run-time requirements per NFPA 110. NFPA 110 sets minimum on-site fuel run-time requirements tied to the Emergency Power Supply System (EPSS) classification (discussed below) — emergency and standby power systems must have fuel supply adequate for their required class-and-type designation without refueling, which is a primary driver of bulk tank sizing for critical facilities such as hospitals.
• Fuel quality management. Diesel fuel degrades over long storage periods (microbial growth, water contamination, fuel gelling in cold climates), so system design for facilities with rarely-run standby generators should include fuel polishing/filtration systems, periodic fuel testing, and tank design features (sloped bottoms, water-draw sumps) that support fuel quality maintenance over years of mostly-idle storage.

NFPA 110 Level 1 vs. Level 2 EPSS Classification

NFPA 110 classifies Emergency Power Supply Systems (EPSS) by Type (maximum time the system can be without power before restoration — the shorter the time, the more demanding the type) and by Level, which describes the criticality of the load being served:

• Level 1 EPSS — required where failure of the equipment to perform could result in loss of human life or serious injuries. This is the classification applied to systems supporting life safety loads in facilities such as hospitals (see our companion Hospital Electrical Design article for how this plays out in a real NEC Article 517 / NFPA 99 essential electrical system). Level 1 systems carry the most stringent installation, testing, and maintenance requirements in NFPA 110.
• Level 2 EPSS — required where failure of the equipment is less critical to human life and safety but is still needed to protect property or maintain business/process continuity. Requirements are somewhat less stringent than Level 1 but still substantial compared to a purely optional standby system.

The Level designation, combined with the Type (restoration time) designation, drives specific NFPA 110 requirements for periodic testing under load, minimum fuel run-time, remote annunciation, and physical installation (such as separation from other building systems), so the classification must be determined early in design since it materially shapes the transfer switch, fuel system, and testing program requirements that follow.

Load Shedding and Priority Load Management

When a generator system's total capacity is less than the full connected building load — a common design choice to control generator size and cost, since backup power often only needs to cover critical/priority loads rather than the entire building — load shedding logic automatically sheds (drops) lower-priority loads if the generator approaches its capacity limit, protecting the generator from overload and preserving power to the highest-priority circuits. Design typically organizes loads into priority tiers (e.g., life safety loads that are never shed, critical process or equipment loads shed only under the most severe capacity constraints, and lowest-priority loads such as non-essential HVAC or convenience loads shed first). Load shedding can be implemented through simple sequential timers that stagger large motor loads on generator start (reducing simultaneous starting inrush) or through more sophisticated generator paralleling switchgear controls that continuously monitor generator loading and automatically shed pre-defined load banks as loading approaches a set threshold, then automatically restore shed loads as capacity becomes available again.

Bringing the System Together

A complete generator system design integrates: an ATS transition mode matched to the actual operational need (open-transition for standard emergency response, closed-transition added only where seamless testing/scheduled transfer is required); a paralleling and synchronization strategy if redundancy or capacity beyond a single unit is required; a fuel system sized to the facility's mandated run-time with day tank buffering and long-term fuel quality management; an NFPA 110 Level and Type classification determined early, since it drives fuel, testing, and installation requirements; and load shedding logic if the generator capacity is intentionally sized below full connected load. Each of these decisions should be made in the context of the generator's already-completed kW sizing from the Generator Sizing Guide, not independently of it.