Investigation and fleet-wide checks
181 lessons from 178 energy incidents, newest first, grouped by technology. Each is taken from the entry it links to, which has what happened and the sources.
Battery storage 13
Confirm that installation follows the manufacturer's clearance and ventilation requirements and that a qualified installer has commissioned the system.
Installers and owners should confirm that battery units are sited, ventilated and protected in line with the manufacturer's instructions and local regulations.
Owners should publish cause findings after a BESS fire, as the owner did here, so other operators can check the same failure mode in their own fleets.
Treat a home battery in the room where a fire starts as a possible source until the investigation is finished, and do not treat suspicion as a confirmed cause.
A single Megapack clamp failure warrants fleet-wide checks on the thermal management system to prevent repeats.
Domestic battery storage systems can be linked to house fires, so installers must confirm systems meet manufacturer instructions and installation standards.
A single residential storage fire should not be read as proof that a particular model failed while the investigation is still open.
When investigating a battery fire, use voltage, current and temperature logs to find the first module to fail, since a failing module typically drops to zero output first.
Keep the recorded cause open when investigators say it is undetermined, even if a most-likely electrical sequence is described.
Owners should keep batteries maintained and protected from damage and heat, and report faults or warnings to the installer, since damaged lithium-ion cells can go into thermal runaway.
Sites without recent detailed safety checks should be shut down promptly after any incident in the fleet.
An earlier fire at another KEPCO frequency regulation ESS did not prevent recurrence, so operators should carry out full fleet diagnostics after any single event.
Sites must complete root cause investigation before returning containers to service, as a second container fire occurred on 21 December 2017.
Solar PV 43
Where a fire is first attributed to a particular asset, the operator should confirm the cause with a documented investigation before drawing conclusions.
Until a fuller report is available, operators should not treat this case as evidence of a particular failure mode.
Where the cause is unknown, the site should be preserved for investigation before repair.
Where a fire destroys on-site solar installations, treat the cause as unconfirmed until a formal investigation reports.
Covering the panels was one tactic used to finish the fire; the cause was still under investigation.
Until police report the cause of a station fire, operators should avoid assumptions and keep records for investigators.
Fire damage to solar panels can destroy evidence, so operators should retain inverter telemetry for later analysis.
Until the cause is known, preserve evidence and treat the event as under investigation rather than assuming a single failure mode.
Installers must match solar unit type to the home's electrical system before fitting, as a mismatch caused two fires.
Homeowners should keep solar PV installations maintained in line with the installer's and manufacturer's guidance.
Involvement of panels is not, on its own, evidence that they caused the fire.
Where the cause is not known, keep the array and electrical equipment available for investigation before repair.
Where the cause is still under investigation, operators should treat animal intrusion as a suspected factor, not a confirmed one.
Complete and share the cause investigation so similar arrays can be checked.
A contained roof fire can still disrupt services, so operators should expect cancelled appointments and a follow-up investigation even when no injuries are reported.
A fire risk at one solar asset can affect several sites at once where equipment or design is shared across a fleet.
Operators benefit from knowing quickly which sites use the same equipment type so checks can be targeted.
Grid and generation equipment ownership must be clearly marked so fault investigations do not mix separate systems.
Keep an accurate inventory of stored equipment to support insurance claims and investigation.
Solar installers or owners should be contacted early to provide technical assistance, and sites should hold current contact details for this purpose.
While the cause is under investigation, do not assume a single component failed, and keep the scene and records available to investigators.
Treat any early link to a panel power unit as provisional until the fire investigation is complete.
Treat an early finding that a fire does not appear suspicious as provisional until the cause and origin investigation is complete.
Suspected causes should be recorded as suspected until the investigation is complete.
Suspected medical emergencies and workplace accidents should be recorded and investigated together until the cause is known, so that health and safety lessons are not lost.
Operators should record damage and service impact promptly after an incident to support investigation and reporting.
Use qualified installers who follow the applicable technical standards, and check that the roof structure can bear the array load.
Until investigators report a cause, operators should not treat either the array or nearby equipment as the established source.
Solar panels removed and reinstalled after roof repairs must be checked by a qualified installer before reconnection.
Operators should share findings from investigations of transformer fires so that other sites can check for the same causes.
Operators and installers should document the fault and its cause so that similar failures can be identified and prevented.
Where the cause is not yet released, operators should preserve the damaged panel and nearby wiring for investigation.
Treat early reports that solar panels appeared to have started a roof fire as unconfirmed until an investigation is complete.
Investigations into solar distribution-board fires should be shared so other operators can check similar equipment.
A same-day belief that a solar panel started a loft fire should stay provisional until a fire investigation reports the cause.
Once investigators name the failed part of the system, check similar rooftop arrays rather than treating the cause as closed on a general solar finding.
A fire at a solar farm requires prompt investigation to identify the cause before further events occur.
Operators should publish the cause and outcome of solar farm fires once investigations close.
Fire and Rescue NSW advises that solar systems be installed by licensed installers and professionally maintained to reduce fire risk.
Reporting the cause of domestic solar fires, once found, helps installers and owners spot common faults.
Investigators should record the installation details of any PV system involved in a roof fire, so that the cause can be established and shared with the industry.
Electric faults in rooftop solar panels can destroy wiring evidence when heat burns it away, so record early observations of the fire origin.
Record the installer of each solar array to allow targeted checks after any fire.
Wind (onshore) 29
Operators should publish or share root cause findings with enough detail to allow other owners to check their own fleets for the same fault.
After any turbine fire, operators should promptly preserve evidence and investigate the cause, so that the findings can be shared with other operators.
Coordination with environmental bodies is essential where turbines stand in shared forest land, to manage temporary closures during investigation.
Share findings with the turbine manufacturer and other operators of the same model, because a failure mode in one fleet may apply to others.
When the operator of a failed turbine is unclear, quick identification is needed to start the investigation.
After hub detachment, the operator took the entire wind farm offline as a precaution during investigation.
Operators should confirm the cause of any turbine fire through a documented investigation and share findings on electrical and mechanical faults with the wider industry.
With no cause reported, operators should still identify the failed unit and preserve debris so the failure can be investigated.
Plan for investigation, safe removal and recycling or disposal of a failed turbine before the site returns to service.
After a turbine fire, the State Fire Investigation Unit inquiry can take several weeks because of scene complexity and investigator safety, so operators should plan for extended downtime.
Manufacturers and operators should agree root-cause investigation responsibilities early, so that findings can be shared across the turbine fleet.
A video title can show that a turbine fire was reported, but missing detail is not evidence that there was no damage or no response.
A claim that schedule pressure is competing with safety should be checked against the work controls, not treated as the cause until it is investigated.
Fallback paths are needed because loss of one communications channel left a large fleet without control for an unknown period.
Early sharing of preliminary findings lets other operators of the same model check their assets without delay.
Two turbines damaged at once at a wind site calls for checks on common mode failure in GE turbines until the investigation rules it out.
Treat the manufacturer's stated cause as a lead to check until an independent investigation is available.
Investigate turbine fires promptly, including events described as rare, and record which turbine model was involved.
Operators should publish a confirmed cause once investigated, so that other sites can act on the findings.
A fatal event on site requires immediate work stoppage until the scene is secured and an external investigation begins.
Handover protocols between police, fire and the operator help scenes pass safely to technical investigation.
A fire at a 20-turbine site with no prior incidents since 2000 still requires full investigation before restarting the turbine.
When the root cause stays unknown, the operator should log the finding and check matching electrical systems across the fleet.
Do not state a cause until an investigation has reported one.
Taking a small part of a plant offline while the cause is investigated can limit further risk.
Operators should record the outcome of the investigation so similar turbines can be checked.
Delivery schedules for a new turbine model should wait on the outcome of any fire investigation.
When root cause analysis fails due to damage, check every similar turbine in the fleet as a precaution.
Early headlines give no cause, so operators must await full investigation before any diagnosis.
Wind (offshore) 9
Where a third party may be liable, operators should record damage evidence promptly and preserve the information needed to support a compensation claim.
Early repeat failures on a first-of-a-kind project should prompt root-cause investigation across the whole fleet, not only the affected units.
Joint investigation and repair plans with the turbine manufacturer must be agreed in advance to speed recovery after an incident.
Offshore wind farms in fishing areas need clear marking and communication to local fleets.
High potential near misses with no injury still need full investigation. Operators should submit and review anonymised cases so teams learn detachment modes seen elsewhere.
Precautionary no-sail zones around the same turbine model in multiple countries must be based on confirmed risk, not assumed fleet-wide failure.
Investigate unplanned movement of heavy-lift crane components even when no one is injured, so the cause is known before further lifting.
When the cause of a Siemens Gamesa turbine fire remains unknown, operators and the supplier must preserve and share site evidence before further work.
A multi-casualty event on an installation vessel should trigger an immediate return to port to preserve evidence.
Grid infrastructure 73
Preserve fault records and equipment evidence after an incident so that the cause can be established.
Utilities should record the time, equipment involved and exposure checks promptly so that the cause can be investigated and confirmed later.
Independent investigations with a short, fixed reporting deadline help identify root causes and recurrence controls after major grid failures.
Ownership boundaries between consumers and distribution operators should be documented so that each incident is investigated by the party responsible for the equipment.
Operators should publish the cause of substation fires once investigated, so that other operators can learn from them.
Treat repeated faults at a single critical substation as a pattern and investigate them together rather than as separate events.
Operators should keep a tracked record of investigation recommendations and confirm their implementation.
A preliminary electrical-fault finding should be confirmed by investigation before it is treated as the cause.
Where the cause is not known when the fire is out, an investigation should start as soon as the scene is safe.
An ordered investigation should establish the cause before operators draw conclusions about the failure.
Do not state a cause until the investigation reports one.
Treat a suspected internal fault as unconfirmed until the investigation is complete, and do not return the circuit to service on that assumption alone.
Investigate the circumstances and causes of a fatal transformer explosion and share findings other operators can use.
Until investigators say otherwise, treat both equipment failure and nearby vehicle activity as open questions and preserve the scene.
Explosion sounds heard before a substation fire require evidence preservation and no assumed cause.
Do not state a cause or a link to other events until it is confirmed by investigators.
A suspected short-circuit should stay marked as unconfirmed until the investigation is finished.
When a pole-mounted transformer fails after a pole fall, treat the cause as unknown until the utility investigation is reported.
Where both a technical fault and a crime remain open, operators should preserve evidence and avoid closing the cause early.
A technical investigation into a transformer fire must continue alongside repairs, so operators should plan for parallel teams from the outset.
If an external cause is possible, operators should record the event and preserve evidence so that investigators can later establish what happened.
Until the investigation reports, the origin of the fire should be treated as unknown.
Do not treat the failed component as the root cause until the investigation is reported.
When a bang is reported but not confirmed, treat the event as unclassified until the official investigation concludes.
Reports of explosions and flickering lights can precede a substation fire, so operators should log and investigate such reports promptly.
Early reports of cause, including attribution to sabotage, should be treated as provisional until investigators confirm the facts.
Operators should avoid presenting unconfirmed links between incidents as fact, and should state clearly what is still being investigated.
Investigate mechanical faults in generator and transformer equipment promptly and share findings with other operators.
Early, transparent investigation of substation explosions helps other operators check similar equipment and failure modes.
Until investigators report a cause, operators should not treat a substation fire as explained.
The failed unit should be investigated promptly, because a single affected household can still leave the cause unknown at the time of the fire.
Do not treat suspected sabotage as a confirmed cause until an investigation is complete.
Maintain clear records of equipment status to support post-incident investigation.
Hand the scene to the network operator for damage assessment, and do not treat the cause as known until that investigation is reported.
Clear suspension of work and investigation protocols after containment reduce further risk.
Treat early reports of an explosion as unconfirmed until the investigation is complete.
Where two fires are involved, investigation should cover both the origin and how the substation was affected.
Battery fleets can inject 1.5 GW within seconds, so contracts should require fast response from aggregated assets across multiple optimisers.
Operators should investigate short-circuit causes promptly and share the findings, so that similar transformers can be checked for the same fault.
Until the cause is known, operators should treat the event as under investigation and not assume a single failure mode.
Involve specialist investigators where irregularities are alleged around critical electricity assets.
Treat an initial official finding as preliminary until the full investigation is published.
Until an investigation reports a cause, treat the initiating failure as unknown rather than assuming a single component.
Until the cause is established, operators should treat the event as unexplained and preserve evidence for the investigation.
Where the cause is not known after the initial response, operators should preserve evidence and publish findings when they are established.
Cause investigations should be completed and the findings used to check similar sites.
Where the cause is not yet known, operators should preserve evidence and complete an investigation before drawing conclusions.
Where the cause is unknown, operators should investigate before treating the event as closed.
Evidence should be preserved and a formal inquiry completed before a damaged unit is returned to service, so that the confirmed cause informs repairs and other sites.
Operators should publish a clear cause investigation so that other utilities can learn from substation fire events.
Where early reports disagree on cause, treat the cause as unconfirmed until the investigation is finished.
Early investigation and site control help when deliberate interference is suspected.
Where a distribution fire is reported soon after load-shedding, record the timing and wait for the investigation before treating shedding as the cause.
Post-incident investigations should record the ignition point and failed equipment so that causes can be shared across the sector.
Root cause of a substation transformer fire remains unknown, so operators should complete and share a full post-incident investigation.
Operators should investigate substation fires promptly and publish the root cause so that other sites can check similar equipment.
Operators should investigate substation fires and share the findings, so similar equipment elsewhere can be checked.
Where the cause is not yet known, keep the site attended and investigate before the equipment is returned to service.
Publish investigation findings when they exist, so other operators can see whether similar equipment is involved.
A serious injury and an unexplained transformer incident should be investigated with the relevant authorities before the unit is treated as fit to return.
The newest 60 of 73 shown. The rest are on each grid infrastructure entry.
Data centre 6
Keep the original equipment supplier engaged and contractually responsive for fault investigation and repair.
Treat the cause as unconfirmed until the investigation reports.
Until the cause is known, a reported site gas leak should be treated as an open investigation, not a confirmed equipment failure.
Repeated fires at different sites in the same city are a reason for operators to share what investigations find.
Investigate all electrical incidents promptly to identify and mitigate recurring hazards.
Preserve the failed unit and its fault data so that the vendor and investigators can establish the cause.
Hybrid co-located 5
Operators and investigators should publish confirmed root causes once investigations conclude, so that other operators can address the same failure modes.
A post-event fire watch and removal of the affected cabinet from service allows safe handover even without an active fire.
Share fire-investigation findings with other operators of hybrid solar and storage sites.
Treat an official belief about cause as provisional until an investigation is reported.
Operators of three-and-a-half-year-old systems should review shared investigation findings to assess similar fire risks.
Other 3
A trip on one unit at a captive generating plant can stop an industrial process even when other units remain available.
Separate unverified social media claims from the reported sequence until an investigation is available.
Start a documented cause investigation early and share initial findings where it is safe to do so, so that other operators can check similar work practices.