Electrical
How to prove solar and diesel generator transfer performance
Witness the operating sequence, signed power balance and controller failures that establish solar-diesel transfer performance.
Direct answer
Prove solar and diesel generator transfer performance by witnessing each approved operating sequence, including source loss, diesel supply, PV admission or exclusion, load changes and return to normal, with time-aligned power and breaker records. Test minimum-load control, reverse-power protection and controller failures separately, then demonstrate their coordinated response. Use the installed engine limits, approved protection settings and interconnection requirements as the acceptance criteria; a successful synchronization or load-bank run alone does not prove the PV-diesel interface.
On this page
What proves solar and diesel generator transfer performance?Which operating states and transfers must the plant support?Which documents set the acceptance limits?How do you prove reverse-power protection?What minimum generator load should PV control maintain?How do you prove reserve and response to sudden load changes?What measurements and recording intervals capture the transfer?Which transfer events must be witnessed?What should happen when a controller interface fails?Worked example: a 450 kW generator serving a 400 kW loadHow long should testing take, and what belongs in the scope?Field checklist and turnover evidenceWhen should solar-diesel acceptance be repeated?Open the related tool for the calculation, checklist, or packet.
FieldOS by AnvilfieldOffline-first FieldOS for job photos, proof packets, property memory, calculations, teams, billing, and owner-ready reports.What proves solar and diesel generator transfer performance?
Prove solar and diesel generator transfer performance by witnessing each approved operating sequence, including source loss, diesel supply, PV admission or exclusion, load changes and return to normal, with time-aligned power and breaker records. Test minimum-load control, reverse-power protection and controller failures separately, then demonstrate their coordinated response. Use the installed engine limits, approved protection settings and interconnection requirements as the acceptance criteria; a successful synchronization or load-bank run alone does not prove the PV-diesel interface.
DG means diesel generator; PV means photovoltaic generation. Establish measurement boundaries at each generator terminal, the PV connection and the transfer/load bus, including auxiliaries and losses consistently. A standby generator becomes part of a hybrid system only where the approved architecture permits PV participation.
Retain five linked evidence units: the operating states, reverse-power register, minimum-load controls, transfer event record and controller failure matrix. Anvilfield recommends these records to connect decisions to power flow and load continuity.
Acceptance means the required loads meet their interruption and voltage/frequency limits, utility isolation remains effective, engine duty stays within its approved envelope and failures produce the specified fallback. Complete the applicable generator synchronization and standalone generator acceptance first; this witness establishes the interfaces between those accepted components.
Which operating states and transfers must the plant support?
Identify the source that establishes voltage and frequency in every state, and the device that isolates the utility. Grid-following PV needs an established compatible reference and permission to operate in that island; an ordinary grid-interconnection certificate does not establish diesel-island compatibility. NREL/Sandia, December 2022 separates grid-connected operation, transition, island operation and reconnection.
UTILITY + PV | source loss; utility isolation v OPEN TRANSFER (PV blocked) | diesel ready; source path proved v DIESEL ONLY | island permission; controls healthy v DIESEL + PV -- PV loss/curtailment --> DIESEL ONLY | interface failure v APPROVED FALLBACK | fresh data; permissives; controlled ramp +---------------------------------> DIESEL + PV PV excluded by design: remain DIESEL ONLY. Qualified utility return: approved retransfer, then diesel separation and normal PV mode. Grid-forming source lost from any diesel state: PV INHIBITED / DEAD BUS, unless approved storage takes over. Black-start + isolation -> DIESEL ONLY.
| State | Grid-forming source and breaker condition | PV permission | Witness criterion |
|---|---|---|---|
| Utility connected | Utility supplies reference; approved normal path closed | Normal grid settings | Utility interface/export agrees with agreement |
| Open transfer | Normal path open before alternate closes | Block until alternate source and island mode accepted | Record dead interval; no unintended source overlap |
| Diesel-only island | Diesel bus energized; utility isolation proven | Initially blocked | Correct V/f, signed kW and load path before PV admission |
| Diesel plus PV | Diesel remains grid forming in this case | Limited by minimum load, reserve and equipment | Per-engine floor respected in settled operation; transients inside approved envelope |
| Diesel supply with PV excluded | Approved transfer arrangement keeps PV separated or inhibited | Blocked throughout diesel supply | Prove exclusion, absence of PV backfeed and controlled grid readmission |
| Closed retransfer | Utility and generator overlap only when authorized | Mode-specific | Sync permissive, overlap duration and export within agreement |
| Interface fallback | Source path determined by fault sequence | Actual configured fallback, not assumed zero | Measured PV kW and diesel load agree with fault design |
| Grid-forming source lost | No reference unless approved BESS takes over | Inhibit incompatible grid-following PV | No unintended energized island; documented recovery |
Open transition, closed transition and permanent island operation
Open transition deliberately interrupts supply; measure the dead interval at the load bus. Closed transition permits a controlled source overlap and requires the approved synchronization, overlap and export limits. Use the ATS commissioning record to establish actual contact sequence, rather than treating an engine-running indication as restored load power.
A permanent island omits utility-transfer tests but retains PV loss, load rejection, grid-forming source loss and recovery. Battery grid forming adds a diesel-off state bounded by state of charge and converter limits. PV admission to an energized diesel bus is distinct from load transfer between sources.
Which documents set the acceptance limits?
Before energized testing, issue an acceptance schedule naming each measuring point, allowed band/pickup, duration, tolerance, output, abort action, document revision and approving party. Cover engine floor/recovery duty, step response, voltage/frequency recovery, load interruption, PV response, fallback timeout, overlap and export.
| Authority / edition checked | What it establishes | What supplies plant acceptance limits |
|---|---|---|
| IEEE 2030.8-2018, publisher page active on review date | Testing and performance comparison of microgrid controller functions | Contract test specification and licensed applicable provisions; IEEE 2030.8-2018 |
| IEEE 1547-2018 and 1547a-2020, current publisher/committee listings | DER interface with the Area Electric Power System (Area EPS) | Utility agreement, adopted requirements, installed DER settings; NREL, 2022, IEEE 1547-2018, IEEE 1547 series register, April 2026 |
| Cummins T-030, A040S369 Issue 9, February 2015 | Generator application and ratings | Installed engine/model/duty documentation; Cummins Issue 9 |
| SMA FSC20-BA-xx-20 v2.0 | That controller's settings and coordination | Approved protection study and verified installed revision; SMA v2.0 |
| DEIF ASC-4 4189341077D EN | Commissioning, communications, reserve and downstream interfaces | Installed device manual, interface map and approved action logic; DEIF rev D |
| Power and Water handbook, second edition, hosted 2019 | Operating experience and integration calculations | Actual engine limits and site transient study; Handbook, 2nd ed. |
NREL/Sandia, December 2022 explains that IEEE 1547-2018 Clause 8.1 requires DER to detect an unintentional Area EPS island and cease energizing the Area EPS and trip within 2 seconds, with an alternative of forming an intentional island isolated from the Area EPS. That interval is not a diesel-start requirement. IEEE 1547.1-2020 equipment-conformance evidence and IEEE 2030.8-2018 controller-function testing address different questions; the installed plant still needs its witnessed sequence.
Put generator model and duty rating, site derating, inverter firmware, controller version, meter scaling, approved one-line, utility agreement and load ride-through requirements on the witness-pack cover. Identify the neutral and grounding arrangement in both source modes. Retain applicable emergency-power acceptance as its own obligation.
Resolve conflicting requirements through the approved sequence and protection coordination study. Selecting the shortest timer can trip the generator before PV control acts, defeating selectivity. Publisher edition status establishes the document being cited; jurisdictional adoption and the utility agreement establish its application.
How do you prove reverse-power protection?
Use positive generator kW for delivery to the bus and negative kW for import into the generator, then verify the installed meter convention. Reverse kVAR or low power factor alone is not reverse active power. A positive generator output below its minimum operating floor is underloading, even when reverse-power protection has not picked up.
SMA's Fuel Save Controller 2.0 manual, FSC20-BA-xx-20 version 2.0, requires its reverse-power protection to act before the generator's protection, with generator protection remaining active. Its defaults below are reference values for that revision; enter installed settings and measured results alongside them. SMA FSC manual v2.0, pp72 to 73
| Function | Source value and basis | Delay | Intended evidence | Acceptance distinction | Source |
|---|---|---|---|---|---|
| FSC device reverse-power error | -5% of nominal generator active power, default | 0.1 s default | Inject directional condition; capture error and downstream action | Error timer is not end-to-end clearing time | SMA v2.0 p72 |
| FSC total reverse-power error | -5%, total-function default; verify the installed aggregate nominal kW denominator and online-set selection | 0.1 s default | Confirm aggregated input/rating basis and response | Total function does not replace per-set checks | SMA v2.0 p72 |
| FSC device/total warning | +5%, default; retain positive sign | 2 s default | Prove warning at low positive output using configured direction | Warning is not a negative-power trip | SMA v2.0 p72 |
| Generator protective function | Approved installed generator relay setting | Approved installed relay delay | Independent injection/output-path test and coordination trace | Keep generator protection active; FSC acts first in this SMA scheme | SMA v2.0 p73 |
| Meter polarity and scaling | Forward generator delivery reads positive with the chosen convention | Not a protection timer | Per-phase analyzer comparison under known load | Investigate wrong CT phase/polarity before changing protection | DEIF, 2023 |
Add equipment ID/firmware, CT/VT ratios, approved settings, measured pickup/time, uncertainty, output destination, file ID and witness. Record percentage denominators in kW and online-set selection; verify aggregate calculations when generators join or leave.
Prove polarity and scaling before changing the relay
DEIF reverse-power troubleshooting, 2023 identifies CT polarity, speed regulation and load-ramp behavior as distinct reverse-power troubleshooting paths. Under known forward load, compare phase kW, CT/VT ratios and phase association with an independent analyzer. Correct wiring through an approved isolated test arrangement; never raise a protection threshold to conceal a sign error. Never open an energized CT secondary: OSHA 29 CFR 1910.269(w)(2), for covered generation work, expressly prohibits it and requires bridging where primary deenergization is unavailable.
Separate injection testing from controlled plant response
Anvilfield recommends approved secondary injection or controller emulation for directional pickup, timing, alarm mapping and output-path proof. Test operation beyond pickup, restraint on the non-operating side, dropout and timer reset behavior. Use the installed function tolerance and test-set uncertainty, keeping the engine protected from unintended motoring. Controller-input emulation must be supplemented by independent relay injection and output-path proof; it does not test CT wiring or the generator relay.
Then witness bounded load steps and actual PV curtailment inside the approved engine and load-bank test envelope. Capture detection, command, inverter kW and protective action on one time base. Stop at the approved abort limits; injection proves protection behavior, while the energized test proves the downstream electrical response.
What minimum generator load should PV control maintain?
Cummins' T-030 Application Manual, A040S369 Issue 9, February 2015, recommends operating a generator set at a minimum of 30% of its nameplate rating, while the installed engine's approved duty and low-load conditions determine the acceptance floor. Cummins T-030, Issue 9 gives general application guidance; Power and Water's original handbook, §7.2.2 describes legacy-engine floors of 40% to 60%; its second edition reports the lower-load installation below.
| Source/application | Minimum-load value | Other condition | Control or witness consequence |
|---|---|---|---|
| Cummins general application guidance, T-030 02-2015 §2.4 | Recommended at least 30% of nameplate | Match duty/rating and installed engine guidance | For a 450 kW rating, reference floor is 135 kW; Anvilfield estimate |
| SMA FSC v2.0 parameter GenPwrAtMinPc | 30% default; adjustable 0% to 100% | System minimum-load setpoint; separate underload deadband/timer | Record approved engine floor, permitted excursion duration and actual curtailment |
| Power and Water original handbook, §7.2.2, p32 | Historical 40% to 60% | Engine-by-engine limit varies with age, RPM, governor and design | Do not impose 30% on a legacy engine requiring more load |
| Power and Water reported 230 kW QSL9 low-load installation | 10%, calculated 23 kW | Daily loading above 60%, calculated above 138 kW; duration not given in case | Obtain the installed engine's complete high-load recovery requirement |
| DEIF ASC Plant Management product description, checked 2026-11-05 | Configured engine minimum, no universal percentage | Described minimum-load constraint applies in island mode | Verify correct mode selects the intended control |
| DEIF ASC-4 commissioning rev D | Two minimum-load settings for different engine types | Apply correct engine selection | Prove both configured engine groups if fitted |
Sources: SMA v2.0, pp70 to 71, Power and Water, 2nd ed., pp20 to 22, DEIF ASC Plant Management and ASC-4 rev D, pp48 to 49. Keep the engine floor, controller operating margin, underload alarm deadband and alarm delay separate. An allowed brief excursion does not authorize sustained low-load operation. Power and Water's second edition, p22, specifies daily high loading for its QSL9 case without a duration; the installed duty documentation must supply that recovery requirement.
Calculate the PV ceiling at the measured bus
Subtract all online engines' combined minimum output from demand for the PV ceiling, then apply available PV, reserve and equipment limits. Use active-power ratings for installed duty and ambient conditions; never subtract kVA from kW.
Anvilfield calculation method, no-storage common bus: when P_load ≥ ΣP_min,i, P_PV,ceiling = P_load − ΣP_min,i. When P_load < ΣP_min,i, no feasible steady operating point satisfies those floors.
Another online generator adds its minimum contribution and reduces PV headroom. Below the combined floor, change dispatch, use approved lower-load duty, charge available storage or add suitable load. Charging requires both charge-power capability and energy headroom.
How do you prove reserve and response to sudden load changes?
PV loss demands an upward diesel response; load loss demands rapid reduction of generation. A steady PV ceiling proves neither transient. Power and Water, second edition uses enough diesel capacity to cover 100% of PV output in its medium-contribution approach; DEIF ASC-4, revision D describes dynamic reserve as a configurable percentage of actual PV output, with 10% as a setting example rather than a universal adequate reserve.
Prove available kW headroom and permitted engine step response separately. Size the witnessed PV-loss event to the approved credible loss, and include any specified coincident load pickup; a selectable reserve percentage alone establishes neither. Record frequency nadir and peak, voltage dip and rise, recovery and lost load during the approved step. Cummins T-030, Issue 9 distinguishes rating and application conditions: cold, preloaded and ambient-derated machines can accept different steps.
Anvilfield recommends a response budget covering measurement/filter delay, controller scan and decision, communications, inverter ramp and independent trip or physical opening. A fast pickup setting does not accelerate a slow setpoint channel. Use measured installed delays to establish margin and bounded live steps; do not combine timing defaults from different products into a coordination claim.
Check generator current, kVA and excitation limits alongside inverter current/reactive-power capability, load power factor and phase imbalance. Those limits can bind before active power reaches its ceiling. Witness material motor or reactive-load steps included in the accepted duty; a resistive bank does not reproduce them.
Multiple engines, storage and weather change the envelope
For multiple-generator switchgear, record each signed kW and remaining reserve after a set is lost. Apply BESS commissioning evidence to charge/discharge power, state of charge, temperature, availability and grid-forming mode. Repeat high-PV/low-load and PV-loss boundaries with irradiance, temperature and derating recorded.
Insufficient sunlight leaves the high-PV physical-response test outstanding. An agreed emulator, such as the capability described for DEIF ASC 150 Solar, can establish logic behavior while a later energized witness establishes actual inverter response.
What measurements and recording intervals capture the transfer?
Capture signed per-generator and PV kW, kVAR, bus voltage/frequency, source/load current, commands, breaker auxiliary contacts, controller mode, communications health and alarms. Add irradiance and storage power/state of charge where fitted. Retain native waveform and event files so a power-quality investigation can reconstruct the disturbance.
For a 0.1-second setting, Anvilfield recommends a derived-power/event interval no slower than 10 ms and verified clock alignment within 1 ms for this witness setup. Ten logged intervals across the timer do not guarantee ten independent measurements: overlapping averaging windows, filter delay and trigger placement affect the result. Use the injection test set’s verified timing channel for precise relay timing. Derived-power samples are separate from raw voltage/current waveform sampling, which must resolve the specified disturbance and match the analyzer bandwidth.
| Quantity | Start and finish evidence | Acceptance comparison |
|---|---|---|
| Load interruption | Load-bus supply leaves its acceptable band to acceptable alternate supply | Load/EPSS requirement; engine-running indication alone is insufficient |
| Reverse-power function operation | Applied directional power crosses approved pickup to function output/contact | Installed pickup, delay, tolerance and reset behavior |
| PV curtailment response | Disturbance measurement, controller command, then measured PV kW reaches required band | End-to-end response budget and permitted engine transient |
| Failure/fallback | Last valid data or injected fault to detection, then actual fallback output | Configured timeout plus physical response, including last-value hold |
| Return/overlap | Source qualifies, both actual source paths overlap if permitted, then separation occurs | Approved qualification, overlap/export and re-entry settings |
Disclose voltage/current acquisition, RMS and power averaging windows, scan rates, timestamp resolution, time source and calibration. A one-second historian or minute average cannot prove a subsecond event. Capture pre-event history and the full recovery, sizing the record for the slowest relevant timeout. Record load-side UPS bypass, contactor dropout and process restart where these determine whether the served equipment survived.
Quantify instrument timing uncertainty, filtering delay and clock skew against the acceptance tolerance. When that uncertainty crosses the pass boundary, improve the measurement or repeat the test.
Which transfer events must be witnessed?
Run each applicable transition from a declared starting condition, with approved abort actions ready. Anvilfield recommends at least one recorded run per applicable transition and repeats of every corrected failure and affected configuration; the contract can require more. Use simulation or an isolated test facility for rejected permissives and failed-opening cases; never bridge live sources or defeat utility isolation to demonstrate an interlock.
| ID / trigger | Required recorded sequence | Measurement and interval | Pass evidence |
|---|---|---|---|
| TR-01 normal source loss | Loss detected, utility isolated, diesel ready, alternate path closed, load restored | Load-bus V/f and contacts; proposed 10 ms event interval, 1 ms clock alignment | Interruption and recovery meet declared load requirement; correct isolation |
| TR-02 PV admission | Permission, setpoint, acknowledgement, actual PV increase, diesel reduction | Signed per-set kW and PV kW; trend plus event capture | Engine floor and reserve remain satisfied |
| TR-03 load rejection | Load falls, diesel kW falls, controller responds, actual PV reduces | Full trigger capture; report kW minimum and duration below floor | Approved transient envelope and protection coordination met |
| TR-04 PV loss | PV output falls, diesel accepts load, V/f recovers | PV and diesel kW, voltage minimum, frequency nadir, recovery time | Required loads retained and generator step acceptance met |
| TR-05 interface failure | Channel loss, detection, fallback command/output, electrical fallback | Communication timestamps plus actual PV kW | Specified failure action occurs within configured/approved timeout and response budget |
| TR-06 utility return | Normal source qualified, open or authorized closed transfer, diesel separated | Load V/f, source contacts, overlap where fitted, signed utility kW | Correct sequence, permitted overlap/export, acceptable interruption |
| TR-07 restart after fault | Health restored, permissives checked, controlled PV ramp, auto mode restored | Settings identity, PV/DG kW and alarms | No uncontrolled restart or unintended power step |
| TR-08 rejected transition | Failed start, missing isolation, invalid sync or lost return-source qualification | Input stimulus, permissive, close output, actual contacts and bus response | Unpermitted close blocked; specified fallback/separation and alarm occur |
| TR-09 PV excluded / night run | Diesel supplies load with PV unavailable or inhibited | PV breaker/inhibit state, measured PV kW, load bus and grid-return sequence | Approved diesel-only duty and controlled PV readmission proven |
Use a cause-and-effect record the witness can reconstruct
Every executed row needs test ID, date, UTC/local time basis, starting load/PV/per-engine kW, online equipment, weather, settings file, criterion document and revision, measured extrema and delays, disposition, native file, witness and exception. Record trigger, command, contact change and electrical response separately; acknowledgement is not breaker motion or delivered curtailment.
Prove the rejected-transition cases in TR-08 and failed-overlap logic through the approved test facility. Record critical-load consequences and recovery; the night run establishes diesel-only duty, while available sunlight is needed to prove PV exclusion against attempted production.
What should happen when a controller interface fails?
Name each failed channel's response in the approved cause-and-effect schedule. PV commands, measurement return and generator-management communications are separate fault paths. Anvilfield recommends this matrix and independent recording through controller power loss.
| Failure injected using approved test facility | Detection evidence | Required response to prove | Record / failure indication |
|---|---|---|---|
| PV setpoint link interrupted | Channel timeout and inverter-side status | Installed fallback actually limits PV to the approved value | Record resulting kW and time; reference percentage uses inverter nominal AC power |
| Generator/mains measurement stops updating | Stale-data age or communication/measurement alarm | Bad input invalidates control and invokes the approved fallback | A frozen plausible kW value must not be accepted as fresh data |
| Broadcast setpoints not acted on | Independent measured PV response or configured returned monitoring | Independent protective action if curtailment fails | No handshake alone; DEIF monitoring alarm unavailable without monitoring return |
| Controller loses auxiliary power | Power-loss contact/heartbeat and inverter behavior | Hardware/inverter fallback established by the design | Capture controller shutdown and actual PV response on independent recorder |
| Generator management network lost | Network fault and online-set status quality | Block unsafe dispatch/PV changes; execute configured fallback | Capture per-engine kW, reserve and actual mode |
| PV breaker commanded open but stays closed | Command/contact mismatch, electrical output | Failed-open action reaches the independent designed isolation/control path | DEIF lists PVB open failure 2160 and position failure 2180 |
| Contradictory or missing breaker feedback | Auxiliary contact discrepancy | Block unproven source closure or PV admission | DEIF lists PVB close failure 2170 and position failure 2180 |
| Inverter retains old setpoint | Requested-versus-measured PV disagreement | Controller detects failure or independent protection limits hazard | Register command, timeout and reduction actually achieved |
| Communications restored | Health indication and valid data age | Controlled restart with matched fallback parameters | SMA warns mismatched fallback settings can cause power fluctuations |
| Grid-forming source trips | Source breaker state, bus V/f, remaining sources | PV inhibited or approved storage takeover succeeds | Exercise from diesel-plus-PV and fallback states, with recovery recorded |
| Utility isolation or overlap clearing fails | Command, real contacts and utility/bus measurements | Block island PV admission or initiate specified independent separation | Failed opening cannot be overridden by a controller mode label |
| Inverter/controller reboots or remote setpoint overrides local control | Reboot/authority status, retained parameters, actual PV ramp | Correct control owner, valid permissives and bounded restart | Record precedence and as-left remote/local authority |
SMA's Fuel Save Controller 2.0 manual, version 2.0, specifies FallbackPc at 10% of inverter nominal power by default, so communications-loss acceptance must verify the installed inverter output limit, available sunlight and actual production. SMA FSC manual v2.0 gives FallbackPc a 1% to 100% range and requires controller/inverter fallback settings to agree. Falling sunlight can conceal an ineffective limit. When the approved response needs zero PV, prove the supported stop/block or isolation path.
DEIF ASC-4, revision D recommends a 1 to 2 second transmit rate for its described serial interface and advises against less than 1 second without specific advice. SMA version 2.0 gives a 0.2-second cyclic-setpoint default. These are interface settings from different revisions and products, not guaranteed inverter-response times.
Broadcast commands require independent response evidence
DEIF ASC-4 Commissioning Guidelines, revision D, p42 requires returned monitoring data for its broadcast PV monitoring alarm. That alarm is unavailable where the inverter manager returns no inverter data; one-way broadcast supplies no handshake proving response.
Where supported, prove each inverter responds before selecting final broadcast operation, then compare measured PV kW with commands and test independent failure action. Configure downstream action for alarm-only outputs. Restore fresh data, correct control authority and a bounded restart.
| Symptom | Discriminating evidence | Resulting action |
|---|---|---|
| Negative phase kW under known forward loading | Phase association, CT direction and independent analyzer disagree with controller | Correct isolated wiring/scaling, then repeat polarity and directional tests |
| Diesel kW falls when load is shed | Independent signed kW confirms it; PV stays high until after command delay | Reassess margin, curtailment timing and independent action |
| One generator imports while total diesel exports | Per-engine traces, governor/load-share commands and ratings | Correct sharing/control coordination; repeat per-set protection proof |
| kW oscillates during operation or recovery | Time-aligned setpoints, fallback settings, ramps and competing control authority | Resolve the demonstrated controller/governor interaction |
| Flat PV output despite changing command | Data age, irradiance/available power and independent PV measurement | Separate stale data, command nonresponse and normal resource limitation |
| Acceptable kW but voltage/current failure | kVAR, phase current, power factor, excitation and inverter limit flags | Correct the reactive/phase-limit problem and repeat the actual load duty |
Worked example: a 450 kW generator serving a 400 kW load
Anvilfield estimates from the Power and Water handbook's second-edition p33 inputs put a 450 kW generator supplying 180 kW at 40% load, permitting 220 kW PV at a 400 kW common-bus load and that engine floor, before losses and other operating limits. The source graphic labels the minimum load 45%; its kW values give 40%. A true 45% floor changes the permitted PV output as shown below. Power and Water, second edition
| Quantity or event | Inputs | Result |
|---|---|---|
| Published operating point | Generator 450 kW; load 400 kW; PV 220 kW | Diesel 180 kW |
| Diesel loading | 180 / 450 | 40% |
| Available rated headroom at that point | 450 minus 180 | 270 kW |
| Full PV loss | PV 220 to 0 kW; load stays 400 kW | Diesel target 400 kW; load step 220 kW; headroom left 50 kW |
| PV ceiling at 40% floor | 400 minus 0.40 × 450 | 220 kW |
| PV ceiling at a true 45% floor | Floor 202.5 kW; load 400 kW | 197.5 kW |
| PV ceiling using 30% reference floor | Floor 135 kW; load 400 kW | 265 kW before other limits |
| 150 kW load rejection | New load 250 kW; initially PV 220 kW | Diesel balance 30 kW; new PV ceiling 70 kW at 180 kW floor |
| Required curtailment after that rejection | Initial PV 220 kW minus target 70 kW | 150 kW |
| 250 kW load rejection | New load 150 kW; initially PV 220 kW | Algebraic diesel balance -70 kW |
| Same load after PV fully stops | Load 150 kW; PV zero; adopted diesel floor 180 kW | Diesel 150 kW, 30 kW below floor; no feasible steady PV setting |
| -5% reference threshold | 450 × -0.05 | -22.5 kW |
| 10% fallback with selected 300 kW aggregate inverter nominal AC active power | 300 × 0.10 | 30 kW PV ceiling; at that output and 150 kW load, diesel target 120 kW, 60 kW below floor |
| Two equal online engines | Two 450 kW engines, each 40% floor; 400 kW load | Total floor 360 kW; PV ceiling 40 kW |
The smaller load rejection requires PV to fall by 150 kW to restore the selected floor. The larger rejection produces an algebraic reverse-power balance and belongs in simulation/injection planning; derive bounded energized steps from the actual transient envelope. At the lowest-load condition, zero PV still needs 30 kW of added demand or charging to reach the adopted floor; the reference fallback needs 60 kW. Record time below floor, minimum signed kW and recovery against approved limits.
Method: Anvilfield estimates use the handbook operating point and selected load-step, inverter-capacity and two-engine extensions shown in the table. Subtract PV from load for the algebraic diesel balance, then compare each online engine with its floor. These balances guide test selection; measured waveforms establish the rotating machine response. The no-storage, common-bus calculation omits separately unmeasured losses and uses an available 450 kW rating without site derating; reserve, current and measured dynamic response remain additional acceptance checks.
How long should testing take, and what belongs in the scope?
Anvilfield estimates two 8-hour field days plus 8 lead hours for preparation/reporting for one generator and one PV-controller interface, totaling 56 person-hours. This assumes completed standalone generator acceptance, an approved sequence, accessible test facilities and coordinated owner/OEM attendance.
| Scope item | Person-hours | Deliverable or scheduling dependency |
|---|---|---|
| Commissioning lead, field | 16 | Two field days; signed event sheets and dispositions |
| Electrical test technician | 16 | Instruments, polarity, injection and event capture |
| PV/controller specialist | 8 | Interface faults, curtailment and recovery |
| Generator specialist | 8 | Engine limits, load response and protection coordination |
| Lead preparation/reporting | 8 | Approved scripts, final configurations and turnover pack |
| Corrected-case retest allowance | 8 additional | Four hours each for lead and technician; add affected specialist time |
Method: Anvilfield estimates dated November 5, 2026 sum the selected task hours above. Day one covers instruments, settings, injection and communications; day two covers plant events and restoration. Price the hours with project labor rates, then add instrument/load-bank rental, fuel, travel, outage arrangements and utility attendance. Extra engines, inverter families, storage and deferred sunlight tests add tasks.
Contractors planning repeat service visits can use PorchRocket Asset Search monitoring to track new, removed or altered visible generator assets with dated views.
Field checklist and turnover evidence
For US workplace maintenance covered by Subpart S, OSHA 29 CFR 1910.333 requires deenergizing exposed parts unless the stated exceptions apply, lockout/tagout and qualified-person verification of deenergization, including backfeed. Only qualified persons work on exposed energized parts. Establish the switching plan, source isolation and test connections before plant operation; a controller stop command is not electrical isolation.
The commissioning lead records results, equipment representatives support their interfaces, and the owner witnesses operational outcomes. Utility attendance follows the interconnection agreement; contractual acceptance follows the named project authority. Use this checklist alongside the signed event sheets, not as a replacement for measured results.
Deliver native traces, final configuration backups, firmware versions, protection-study references, point maps, calibration records and exception closeouts. Link the hybrid evidence to generator turnover and the integrated systems test. Mark failed, aborted and untested states explicitly, with the required retest and final automatic-mode restoration recorded.
Want this checklist to run itself on every job — with photo proof and a signed record crews can hand the customer? That's FieldOS.
When should solar-diesel acceptance be repeated?
Anvilfield recommends interface regression tests after controller/inverter firmware changes, protection-setting changes, CT or meter replacement, breaker-control work, additional PV/engines/storage or a transfer failure. Repeat the affected transitions and common fallback paths with the changed configuration identified. Retain the previous accepted files to establish exactly what changed.
At each planned generator exercise, review hybrid alarms and minimum-load/curtailment trends when PV participates. Set the full periodic interface-witness interval through the equipment instructions, contract, utility requirements and observed faults. Preserve engine and emergency-system exercise requirements separately.
Keep the accepted control files and event records with the PV O&M baseline.
For a specific setting or calculation, include its table fragment and source edition.
Reviewed by Anvilfield: 2026-11-05. Update log: 2026-11-05: Article review: corrected handbook-edition attribution, clarified fallback limits and fault-test safety, and added effort and load-side evidence details.
How to cite this page: Anvilfield, How to prove solar and diesel generator transfer performance, 2026, https://anvilfield.com/field-guides/electrical/solar-diesel-transfer-acceptance/
FAQ
How do you accept solar and diesel transfer, reverse power and minimum load?
Witness each applicable state transition, the signed power balance, protection operation and controller-failure response against the approved settings.
How is a solar DG reverse-power test performed?
Prove directional pickup and timing by approved injection, then witness bounded plant events that verify the actual PV-control response.
How does PV generator minimum-load control work?
It curtails PV so each online diesel retains its approved minimum contribution. If demand is below the combined engine floor with PV off, change dispatch or add approved demand or available charging.
Who should witness a solar-diesel transfer test?
The commissioning lead and owner's designated witness record the outcome, with equipment and utility participation required by the contract and interconnection agreement.
Must every diesel run above 30% load?
Use the installed engine’s approved floor and recovery duty. The Cummins general recommendation, historical engine limits and the Power and Water low-load case differ in scope; the minimum-load comparison identifies those differences.
Does a zero-export controller protect the generator from reverse power?
Measure each generator separately. Zero export at the utility connection can coexist with one engine importing power while another exports; per-engine protection and load-sharing tests are still required.
Is a 0.1-second reverse-power timer the transfer time?
No. A reverse-power timer measures protection-function operation; source transfer, physical opening and PV reduction have separate start and finish events. Use the response-timing table to compare each interval with its own limit.
Does loss of communication always shut down the PV?
No. The SMA manual documents a nonzero default fallback. Verify the installed inverter output, timeout and recovery against the lowest-load case in the controller failure matrix.
Can a load bank alone prove hybrid acceptance?
A load bank provides controlled load. Witness PV admission, curtailment, interface failures and actual load continuity as well; include required reactive or motor duty that a resistive bank cannot reproduce.
How often should the solar-diesel interface be retested?
Anvilfield recommends retesting after relevant hardware, software or settings changes and transfer failures. The site maintenance program sets periodic testing from equipment instructions, utility requirements, contract and operating history.
Can ordinary grid-tied solar stay on when the diesel generator starts?
Only where the approved transfer arrangement and inverter controls support that diesel island; otherwise prove that PV remains excluded during generator supply and returns under the approved grid sequence.
What if site demand is below the generator minimum load with PV off?
No steady PV setting can satisfy that floor; add approved demand or charging, change online engine dispatch, or use the engine's approved lower-load duty.
People also ask
Codes cited in this guide
This guide is written and reviewed against the published standards below. Always confirm the current adopted edition with the authority having jurisdiction.
Sources checked
- Power and Water Corporation, Solar/Diesel Mini-Grid Handbook, original benchmark edition AF-AV reopened with curl 2026-11-05, HTTP 200. Original edition, §7.2.2 p32: legacy 40% to 60% minimum-load range.
- SMA Solar Technology AG, Fuel Save Controller 2.0 User Manual, FSC20-BA-xx-20, 100590-00.02, version 2.0 AF-AV reopened with curl 2026-11-05, HTTP 200. Verified v2.0 English pp70 to 73 and 94 to 95: minimum load, signed reverse-power defaults, coordination, cyclic commands and fallback. Historical revision; installed version governs.
- Cummins Power Generation, Application Manual T-030, Liquid Cooled Generator Sets, A040S369 Issue 9, 02-2015 AF-AV reopened with curl 2026-11-05, HTTP 200. Internal title identifies A040S369 Issue 9, 02-2015; §2.4 p9 recommends minimum 30% loading. The URL upload date is not the edition.
- DEIF, ASC Plant Management product description AF-AV reopened with curl 2026-11-05, HTTP 200. Island-mode minimum-load constraint and configurable reserve architecture.
- DEIF, “After synchronization breaker trips on Reverse power protection?” AF-AV reopened with curl 2026-11-05, HTTP 200. Modified 2023-09-18; per-phase kW/CT polarity, frequency, speed and load-ramp troubleshooting.
- NREL and Sandia, Overview of Functional Technical Requirements for Intentional Islands, NREL/TP-5D00-83301, December 2022 AF-AV reopened with curl 2026-11-05, HTTP 200. December 2022 report, p12: IEEE 1547-2018 islanding requirement and isolated intentional-island alternative; also operating stages and conformance-test scope.
- DEIF, ASC 150 Solar product page AF-AV reopened with curl 2026-11-05, HTTP 200. Manufacturer describes application emulation and Ethernet Modbus communication; no guaranteed numeric plant-response time inferred.
- IEEE SA, IEEE 2030.8-2018, Standard for the Testing of Microgrid Controllers AF-AV reopened with curl 2026-11-05, HTTP 200. Publisher lists 2030.8-2018 active, published 2018-08-24; public scope, not full licensed clauses.
- Power and Water Corporation, Solar/Diesel Mini-Grid Handbook, second edition AF-AV reopened with curl 2026-11-05, HTTP 200. Second-edition pp20 to 22: low-load discussion and 230 kW QSL9 case; p33: medium-contribution reserve and 450/400/220/180 kW graphic. The graphic labels 180/450 as 45%; arithmetic is 40%.
- DEIF, ASC-4 Commissioning Guidelines, 4189341077D EN, revision D AF-AV reopened with curl 2026-11-05, HTTP 200. 4189341077D EN, pp18, 40 to 42, 48 to 51: breaker alarms, serial transmit settings, broadcast monitoring, engine floors and reserve. Alarm downstream actions require configuration.
- DEIF, ASC-4 Solar documentation index AF-AV reopened with curl 2026-11-05, HTTP 200. Manufacturer documentation index; provenance only, not a second independent technical authority.
- IEEE SA, IEEE 1547-2018 publisher listing AF-AV reopened with curl 2026-11-05, HTTP 200. Publisher identifies active IEEE 1547-2018 and lists amendment 1547a-2020.
- IEEE Standards Committee 21, IEEE 1547 Series AF-AV reopened with curl 2026-11-05, HTTP 200. Committee table dated April 2026 lists 1547-2018, 1547a-2020 and 1547.1-2020; revision in ballot is distinct from a published edition.
- OSHA, 29 CFR 1910.333, Selection and use of work practices AF-AV checked 2026-11-05. Curl returned 403; reopened successfully through web access. Subpart S scope; paragraphs (a), (b)(2) and (c)(2) support isolation, backfeed checks and qualified-person requirements.
- OSHA, 29 CFR 1910.269(w)(2), Current transformer secondaries AF-AV checked 2026-11-05. Curl to section main page returned 403; web access opened both pages. Paragraph (w)(2) prohibits opening energized CT secondaries; applicability checked against (a)(1), including Subpart S distinction.