String Inverter vs Central Inverter: An Engineering Guide for US Utility-Scale Solar

Part A — String vs Central Inverters at a Glance

Every inverter in a utility-scale solar plant does the same basic job: it converts DC power from the PV array into grid-synchronized AC power. How the plant is divided among inverters changes much more than equipment count. It changes DC and AC cabling, energy yield, availability, maintenance, protection, reactive power capability, harmonic behavior and, critically, how the plant is represented in interconnection studies.

Two architectures, one job

The important point

The choice is not simply "string = small plant, central = large plant." Many of the largest US solar plants use string inverters, and central inverters remain competitive on large, flat, uniform sites. The right choice depends on:

• Plant size and layout: terrain, block shape and distance from array to collection point.
• Losses: DC versus AC cable losses, mismatch losses and transformer losses.
• Maintenance and availability: spare strategy, repair time and access.
• Redundancy: how much generation is lost when a single unit trips.
• MPPT requirements: shading, uneven terrain, mixed orientations and bifacial gain.
• Grid requirements: IEEE 2800-2022 performance, reactive capability at the point of interconnection and ride-through.
• Project economics: equipment, balance of system, construction and lifetime O&M cost.
  

The 34.5 kV collection system and the main power transformer are common to both architectures. The differences are concentrated at the array and the inverter block, but their effects reach the POI.

DC/AC ratio and clipping

Both architectures oversize the DC array relative to inverter AC capacity to raise energy yield and capacity factor:

US utility-scale plants commonly use ratios between about 1.25 and 1.40. Higher ratios increase clipping during peak irradiance but raise output in the morning, evening and on cloudy days. String inverters allow the ratio to be adjusted finely block by block, which is useful on irregular sites. Central inverters apply one ratio across a large array section, which is efficient on uniform layouts. In both cases the ratio interacts with energy storage: clipped energy on DC-coupled hybrid plants can be captured by batteries.

MPPT granularity, mismatch and terrain

The maximum power point tracker finds the operating voltage that extracts the most power from the strings connected to it. When strings under one tracker see different conditions shading, soiling, different tilt on rolling terrain, module degradation or bifacial rear-side irradiance differences the shared operating point cannot be optimal for all of them. This is mismatch loss.

• String inverters track each small group of strings independently, so mismatch is contained within a few strings.
• Central inverters track a large array section at one operating point, so non-uniformity is averaged across hundreds of strings.

On flat, uniform sites with single-axis trackers, the yield difference is usually small. On rolling terrain, sites with partial shading, mixed tracker rows or uneven bifacial conditions, finer MPPT granularity can recover meaningful energy. The benefit must be quantified with a site-specific energy model rather than assumed.

Losses: DC, AC and transformer

Conductor losses in each part of the plant follow the familiar relationship:

At the same power, a 1,500 Vdc circuit carries less current than an 800 Vac three-phase circuit, but the comparison depends on cable lengths and sizes. In practice:

• Central architectures have longer DC feeders from combiners to the inverter station, but very short LV AC runs.
• String architectures have short DC runs but many LV AC circuits from distributed inverters to the transformer.
• Transformer losses depend on the number, size and loading of step-up transformers; string blocks often use more, smaller transformers or multi-winding units.
• Night-time losses: transformer no-load losses and inverter standby consumption accrue every night and matter in lifetime energy and auxiliary load calculations.

A complete loss model compares the full chain from module to POI for both options, including the main power transformer and collector system losses.

Availability and redundancy



When a single inverter trips, the generation lost is proportional to its share of the plant:

For a 200 MWac plant, losing a 350 kVA string inverter removes about 0.18% of output; losing a 4.4 MVA central inverter removes about 2.2%. String plants therefore degrade gracefully, but they have many more units that can fail. Central plants have fewer, larger failures, so repair time and spare-parts strategy dominate availability. A fair comparison uses failure rates, mean time to repair and the replacement strategy for each option over the project life.



Operations and maintenance

Protection and the National Electrical Code

Utility-scale PV in the US is designed to NFPA 70 (National Electrical Code), primarily Article 690, and Article 691 for large-scale PV electric supply stations of 5 MWac and above under engineered design and supervision.

• DC overcurrent: central architectures fuse each string in the combiner box. String inverters with one or two strings per MPPT input often do not need string fuses, because the available backfeed current is below the module's maximum series fuse rating.
• DC arc-fault protection: Article 690 requires arc-fault protection for many PV DC circuits; Article 691 allows engineered alternatives for large-scale plants. Inverter and combiner arc-detection capability should be confirmed early.
• Ground-fault detection: each inverter monitors its own DC array insulation; central inverters monitor a much larger array, which can make insulation faults harder to locate.
• AC protection: string blocks require coordinated LV breakers or fuses for many circuits; central skids rely on integrated LV and MV protection.
• Collector and substation protection: both architectures contribute limited fault current, so feeder and transformer protection must be set for inverter-based resource (IBR) behavior rather than synchronous machine behavior.

Reactive power and voltage control

IEEE 2800-2022 and most US interconnection requirements specify reactive capability at the POI, not at the inverter terminals. A common requirement is the ability to operate at 0.95 power factor leading or lagging at rated active power:

For a 200 MW plant, this is about 66 Mvar at the POI. The inverters must also supply the reactive losses of the step-up transformers, the 34.5 kV collector cables (partly offset by cable charging), and the main power transformer, while staying within their own voltage limits. Both architectures can meet the requirement, but the margin is set differently:

• String plants add capability in small increments by adding inverters, and spread reactive output across hundreds of units.
• Central plants add capability in large steps; a shortfall may require an additional skid, a shunt capacitor bank, or a STATCOM.
• Voltage at the inverter terminals rises during reactive output and can limit capability, particularly at high grid voltage. This must be checked across the full operating range, not only at nominal voltage.

Grid performance: IEEE 2800 and NERC

US transmission-connected plants are increasingly specified to IEEE 2800-2022, which defines performance at the reference point of applicability (usually the POI or point of measurement). NERC's reliability standards for IBRs developed under FERC Order No. 901, including PRC-029-1 for voltage and frequency ride-through, apply to the plant regardless of inverter architecture.

The architecture affects how the plant achieves compliance:

• Ride-through: every inverter in the plant must ride through the same event. A string plant has hundreds of protection relays and firmware instances, so consistent settings and firmware management are critical. Several NERC disturbance reviews have found plant output losses caused by inverter-level settings rather than the grid event itself.
• Momentary cessation: IEEE 2800 effectively prohibits momentary cessation within the continuous and mandatory operation regions, so inverter firmware must be configured accordingly.
• Fast fault current: IBR fault current is limited to roughly 1.0 to 1.5 times rated current and is controlled by firmware. Positive- and negative-sequence injection during unbalanced faults depends on the inverter model.
• Active power–frequency response: primary frequency response and fast frequency response are commanded by the plant controller and executed by each inverter.

How inverter architecture changes the study model

From a grid-study perspective, the inverter type affects how the plant and its controls are represented in tools such as PSS®E and PSCAD™.

Positive-sequence (PSS®E, PSLF) models

US planning models represent a solar plant as one or a few equivalent generators using the WECC generic renewable energy system models:

• REGC: the generator/converter model, representing the inverter current injection and its limits.
• REEC: the electrical control model, representing inverter-level active and reactive control.
• REPC: the plant controller model, representing POI voltage, reactive power and frequency control.

The single-machine equivalent includes an equivalent pad-mount transformer and an equivalent collector system impedance. The equivalent collector impedance is not the sum of the feeder impedances. A common approach weights each collector segment by the square of the number of inverters it carries:

where Zi is the impedance of segment i, Ni is the number of inverters downstream of that segment, and N is the total number of inverters on the feeder. For example, a feeder with six blocks connected by equal segments of impedance z has an equivalent of z × (1² + 2² + … + 6²) ÷ 6² = 2.53 z, not 6 z.

Architecture-specific considerations include:

• String plants: hundreds of inverters are aggregated into one equivalent. Inverter-level controls are usually simple and fast, and the plant controller does most of the voltage regulation. Parameters must reflect the actual inverter firmware, communications latency and the plant controller's dispatch method.
• Central plants: fewer, larger units; aggregation is simpler, and some studies model multiple equivalents by feeder. Inverter-level voltage control may play a larger role.
• Mixed plants: different inverter models or firmware versions in the same plant may need separate equivalent generators.

EMT (PSCAD) models

Electromagnetic transient studies require vendor-specific, validated models that represent the actual control code. These are increasingly required by transmission providers and ISOs/RTOs for weak-grid areas, ride-through verification, control interactions and subsynchronous oscillation screening. For string plants, the vendor model is usually scaled to represent many inverters; aggregation must preserve the plant's dynamic response, phase-locked-loop behavior and current limits.

Short-circuit and harmonic models

• Short circuit: IBRs are represented as controlled current sources or with vendor-specified fault current tables, not with synchronous machine subtransient reactances.
• Harmonics and resonance: the 34.5 kV collector system's cable capacitance, the transformer inductances and the inverters' output filters form resonant circuits. Many string inverters can shift resonance frequencies relative to a few large central units, and inverter control bandwidth affects harmonic damping. IEEE 519 limits at the POI must be verified with a harmonic impedance model.
  

System strength and weak-grid performance



Short-circuit ratio (SCR) at the POI indicates how strong the grid is relative to the plant:

At low SCR, typically below about 3, IBR plants are more prone to control instability, voltage oscillations and ride-through problems. Both architectures can be designed for weak grids, but the inverter's phase-locked loop, current-control tuning and plant controller gains must be verified with EMT studies. Grid-forming inverters, now offered in both string and central formats, are an emerging option for very weak systems.

Worked comparison: a 200 MWac plant at a 230 kV POI

The following illustrative comparison shows how the two architectures differ for the same plant.

Both options can meet the same interconnection requirements. The decision depends on site terrain, energy model results, balance-of-system cost, the owner's O&M model, equipment lead times, and the interconnection study results.

Hybrid plants and energy storage

Many US solar projects now include battery energy storage. Architecture affects the coupling options:

• AC-coupled storage: batteries have their own inverters and connect to the 34.5 kV collector system; this works with either PV architecture.
• DC-coupled storage: DC-DC converters connect batteries to the PV inverter's DC bus. This is common with central inverters and can capture clipped energy, but it ties PV and storage control together.
• Interconnection limits: the plant controller must enforce the POI export limit and coordinate PV and storage dispatch, which must be modeled in the REPC or a hybrid plant controller model.

A decision checklist

1. Site layout: flat and uniform, or rolling terrain with irregular block shapes?
2. Energy yield: site-specific energy model with mismatch, clipping, and DC and AC losses.
3. Balance of system: DC combiners and feeders versus LV AC cabling and more transformers.
4. Grid requirements: IEEE 2800, ISO/RTO and transmission owner requirements, reactive capability at the POI and weak-grid performance.
5. Model availability: validated PSS®E and PSCAD models for the exact inverter and firmware.
6. O&M strategy: in-house crews or OEM service agreement, spare strategy and access.
7. Supply chain: lead times, domestic content and warranty terms.
8. Lifecycle: mid-life replacement, repowering and future storage integration.

How Keentel Engineering helps

Keentel Engineering supports solar, storage and hybrid projects from early design through interconnection and commissioning, with power system studies from 4 kV to 765 kV.

Part C — Case Studies

The following are illustrative scenarios based on conditions typical of US utility-scale solar projects. They show how Keentel Engineering approaches architecture, modeling and grid-performance problems; site details are generalised and figures are rounded.

Case Study 1 — 150 MWac string inverter plant at a weak 138 kV POI

Situation

A developer plans a 150 MWac string inverter plant on rolling terrain with 470 inverters on a 34.5 kV collector system. The 138 kV POI has an SCR of about 2.8 under N-1 conditions. The transmission provider's initial dynamic study shows poorly damped voltage oscillations after nearby faults.

Analysis

Review of the PSS®E model shows two problems. The equivalent collector impedance was calculated as a simple series sum, overstating the impedance by more than double. The plant controller gains were vendor defaults intended for a strong grid. EMT simulation with the vendor PSCAD model confirms that the oscillation is driven by plant controller tuning, not by the inverters.

Engineered solution

Keentel recalculates the collector equivalent using the power-weighted method, retunes the plant controller voltage control gains and time constants, and verifies the settings across N-1 SCR conditions with EMT simulation.

Outcome

The revised model shows well-damped response for all study contingencies, and the corrected settings are carried into the commissioning plan and model validation.

Case Study 2 — 250 MWac central inverter plant: reactive capability at a 230 kV POI

Situation

A 250 MWac central inverter plant on a flat agricultural site connects at 230 kV. The design uses 58 central inverters at 0.95 power factor capability measured at the inverter terminals. The interconnection agreement requires 0.95 power factor at the POI.

Analysis

A reactive capability study shows that, after transformer and collector reactive losses, the plant can deliver only about 70% of the required lagging reactive power at the POI at full output. At high POI voltage, inverter terminal voltage limits reduce capability further.

Engineered solution

Keentel evaluates three options: additional central inverter skids, a switched capacitor bank at 34.5 kV, and a main power transformer tap change. The selected solution combines a revised tap position with two additional skids, which also increase the DC/AC flexibility of the plant.

Outcome

The plant meets the full ±0.95 power factor requirement across the specified POI voltage range, with documented capability curves for the transmission provider.

Case Study 3 — 120 MWac string plant: collector system harmonic resonance at 138 kV

Situation

A 120 MWac string inverter plant with a long underground 34.5 kV collector system connects to a 138 kV system with nearby capacitor banks. The transmission owner requires a harmonic study to IEEE 519 before energization.

Analysis

Frequency scan studies show a parallel resonance near the 11th harmonic formed by collector cable capacitance, the capacitor banks and system inductance. Background voltage distortion at that frequency could be amplified at the POI.

Engineered solution

Keentel models the plant harmonic impedance using inverter vendor data, evaluates detuning options and coordinates inverter active damping settings with the vendor. A study of capacitor bank switching states identifies the worst-case conditions.

Outcome

With the revised inverter settings and a defined capacitor operating procedure, predicted harmonic distortion at the POI is within IEEE 519 limits, and post-energization measurements are planned to confirm performance.

Part D — Technical FAQ: String vs Central Inverters

References and Further Reading

Links were current at publication (October 2026).



• IEEE Std 2800-2022, Standard for Interconnection and Interoperability of Inverter-Based Resources Interconnecting with Associated Transmission Electric Power Systems. IEEE.
• IEEE Std 519-2022, Standard for Harmonic Control in Electric Power Systems. IEEE.
• NFPA 70, National Electrical Code, Articles 690 and 691. National Fire Protection Association.
• NERC Reliability Standard PRC-029-1, Frequency and Voltage Ride-through Requirements for Inverter-Based Resources, and Technical Rationale. North American Electric Reliability Corporation.
• WECC Second Generation Generic Renewable Energy System Models — Summary. Western Electricity Coordinating Council.
• Generic Models for PV Plants. Energy Systems Integration Group.
  

Disclaimer

This article is general technical information for educational purposes. It is not engineering advice for any specific project and does not create a professional relationship. Inverter ratings, unit counts, losses, reactive requirements and other values are illustrative; actual designs must follow the current editions of applicable codes and standards, transmission provider and ISO/RTO requirements, equipment manufacturer data and project-specific engineering analysis, sealed by a licensed professional engineer where required. Verify the current edition of any standard before relying on it.



Case studies are illustrative scenarios based on typical project conditions, not records of specific client projects. Figures are rounded and indicative.

IEEE, NERC, WECC, NFPA, PSS®E, PSCAD™ and other names are the property of their respective owners. Keentel Engineering is not affiliated with or endorsed by these organisations or any software or equipment vendor.

contact@keentelengineering.com  ·  keentelengineering.com  ·  calendly.com/keentel-engineering/15min

State of Florida — Registry No. 36853, KEENTEL LLC, DBA: KEENTEL ENGINEERING

Copyright 1995–2026 Keentel Engineering. All Rights Reserved.