The short version for the O&M desk.
- The maths: on distributed assets the visit is the cheap part. The crane, the vessel and the weather window are what a wrong call costs you.
- The gap: SCADA tells you the turbine tripped or the inverter dropped off. It doesn't show the access track, the cabinet door, the water line or the text on the local display.
- The route: whoever is already at the site captures what your engineer needs, guided step by step from the ground, and the engineer decides parts, skills and crane before anyone drives. Nobody climbs for a video.
- Who this is for: O&M service managers, planners and technical support leads running wind fleets or utility-scale solar portfolios.
The alarm lands at 06:40. Turbine 14 has tripped on a converter fault, the second time this month, and the nearest crew is two and a half hours away. Thursday is the last calm day before a week of 20 m/s wind.
So the question on the board isn't "what's wrong with 14". The question is whether you send two technicians on a guess, whether they take the converter module or the cooling parts, and whether you put a crane on standby for a job you can't yet describe.
That is what renewable energy remote fault diagnosis looks like from the service manager's chair. Less signal processing, more logistics.
SCADA and the OEM portal already narrow it down. They don't tell you what the pad looks like this morning, and that is usually the piece that decides the crane.
This post covers the part that sits between the alarm and the dispatch: what the person already on site can show you, safely, so the drive, the parts and the lift are all decided on evidence. It's an industry post, so the general mechanics of cutting truck rolls and pre-visit assessment are linked rather than repeated.
If you run O&M for a wind fleet or a utility-scale solar portfolio, the weather window is the clock you work to, and a crew that arrives with the wrong module burns the whole window. Venta Capture, a product of VentaVid, gives the person at the site a guided flow on their own phone, no app and no account, and delivers a structured, sealed case to your engineer. Start for free and build the flow for the fault type that costs you the most windy days.
In this post:
Why remote fault diagnosis in renewable energy is a logistics problem
A field-service manager for HVAC or lifts thinks in truck rolls. A renewables O&M manager thinks in truck rolls, weather windows and lifting equipment, and the last two are where the money goes.
On a wind or utility-scale solar site the visit is rarely the expensive decision. The crane, the vessel and the lost window are. Steve Hanawalt at Power Factors puts solar truck roll costs at roughly $125 for a local visit, $400 for a regional one and $1,500 when it needs an overnight, and works through a 500 kW inverter at $0.10/kWh losing about $180 a day. On that example an overnight roll takes eight days to break even.
That maths is manageable. The wind version isn't.
- Cranes: crane equipment and labour make up 30 to 50% of turbine maintenance cost, a single crane runs $10,000 to $50,000 a day, and mobilisation alone is close to half the crane bill on a one-turbine repair, according to Wind Power Engineering. Moving one takes 20 or more truckloads.
- Access: climbs above 300 feet on sites without proper road access, and offshore visits that WorkTrek puts at upwards of $20,000 each.
- Weather: every offshore intervention needs a vessel, a crew and a clearance, and DNV describes repairs as "complex logistical coordination to get specialized vessels, (un)burial spreads and crews, and faultless spare parts on site as fast as possible." Onshore, a high-wind forecast simply closes the tower.
Put those together and a wrong first call doesn't cost you one truck roll. It costs a crane that sits, a crew that comes back with the wrong module, and a calm week you don't get again until the next front passes. The repeat visit is the same failure it is everywhere else; it just carries a much bigger multiplier here.
What SCADA does not show
SCADA is good at its job. It logs the trip, the temperatures, the power curve, the alarm sequence and the converter states, and on many fleets an engineer can reset a nuisance fault from the desk without anyone moving.
The trouble starts when the reset doesn't hold.
A fault usually raises tens or hundreds of alarms in minutes, and the root cause is rarely readable from the alarm list alone. That is the opening finding of a 2022 MDPI paper on SCADA alarm analysis, which describes operators as "overwhelmed by a large amount of supervisory control and data acquisition (SCADA) alarms when faults occur." The data behind the alarms is typically logged as 10-minute averages, which is fine for trending and poor for pinpointing.
Solar has the same shape of problem with fewer alarms and more brands. Across a 3-GW utility-scale portfolio, inverters caused about 80% of non-weather production losses, a kWh Analytics figure quoted by Solar Power World, and every manufacturer names its fault codes differently. Kyle Williams of Pearce Renewables says it plainly in that piece: "The real key to solving a system fast is access to the diagnostic or fault codes."
So the fault code matters, and the code on the local display is often more specific than the one that reaches the portal. But even with the code, the screen at the control centre is blind to a whole category of things:
- Physical state: a chewed cable, a burnt combiner lid, hail damage across a tracker row, a door that no longer latches, a corroded gland, a cabinet vent packed with dust or seed fluff.
- Environment: standing water at the base of the tower or around an inverter pad, ice on the blades, snow against the combiner, vegetation into the string wiring, a flooded access track.
- Human and security: vandalism, theft of cable, an open gate, a contractor who was on site yesterday and "didn't touch anything".
- The local display itself: the exact text, the timestamp, the sub-code and whatever it shows after a manual reset.
- Sound and smell: the noise the site technician heard from the base, the burnt smell at the inverter door.
None of that comes down the SCADA line. All of it decides which parts go on the van and whether a crane is needed at all.
Remote triage for renewables is the discipline of getting that list answered before the dispatch decision, not after the crew arrives. The general case, remote diagnostics without sensors, is covered elsewhere on this blog; here we care about what's specific to a pad in a field.
What the person on site captures, safely
There is nearly always someone closer to the asset than your crew. A site technician on another turbine. The landowner whose field it stands in.
Or the security or mowing contractor at the solar park, the substation operator, and on an OEM-serviced fleet, the OEM technician who is already up-tower on a planned job.
Getting someone to the base was never hard. Getting usable evidence from them was, because "send us some photos" produces a blurry shot of a tower and a voice note.
The organisation decides what gets captured, in what order, with what instruction, and the person on site just follows the steps. That is what guided capture means, and the general method is the same one used for boilers and lifts. What changes in renewables is the safety boundary and the content of the list.
The safety line first
Nobody climbs for a video. Entering a wind turbine is certified work; the GWO Basic Safety Training standard covers working at height, manual handling, fire awareness, first aid and, offshore, sea survival, and a landowner or a security guard holds none of it. On solar, opening an energised inverter or combiner enclosure is qualified electrical work.
So the flow is built in two layers:
- Ground level, exterior only: anyone the operator authorises can capture this. Nothing is opened, nothing is climbed, nothing is touched.
- Up-tower or inside an enclosure: only a certified technician who is already there under their own procedure. The flow is a checklist for them, not permission for anyone else.
A capture flow can carry that instruction in plain words on the first step, and the warning shown before recording is configurable per organisation.
Wind turbine, from the ground
A typical ground-level flow for a tripped turbine asks for:
- Asset identification: scan the turbine number or the QR on the base door, so the case lands against the right asset without anyone typing.
- The controller display: a photo of the base cabinet HMI through the window or on the exterior panel, then a short video reading the fault text, code and time out loud.
- The base and door: cabinet and door exterior, lock state, signs of water, oil, or forced entry.
- The tower and rotor from a safe distance: a slow video of the visible tower, blades and nacelle from the ground, looking for ice, lightning damage, oil streaks, a parked rotor that shouldn't be parked.
- Sound: 20 seconds of audio from the base with the wind noted.
- Access and weather: the track, the crane hardstanding, standing water, snow, and a yes/no on whether a truck could get to the pad today.
- Transformer kiosk exterior: doors, indicators, anything scorched.
Wind turbine, up-tower (certified technician only)
When your own or the OEM's technician is already in the nacelle, the same flow adds the nacelle display, visible leaks at the gearbox and hydraulics, the brake, the converter cabinet indicators, and whatever their procedure already has them opening. The phone stays in their pocket until the work permits it; the flow is there so the record is consistent, not to add a task at height.
Utility-scale solar
For an inverter or string fault the ground-level flow asks for:
- Inverter identification and display: scan the inverter label, then photograph the display and read the fault code and any sub-code aloud.
- Inverter exterior: doors, vents, filters, any scorching, smell noted in a yes/no.
- Combiner boxes on the affected strings: exterior only, lid state, burn marks, water, animal damage.
- The affected rows: a walk-past video of the modules and trackers, hail or storm damage, tracker position, vegetation into the wiring, shading.
- Site conditions: flooding, the access track, gate and fence, evidence of theft or vandalism.
The Solar Builder technicians' guide to inverter troubleshooting lists what a tech wants before dispatch: model and serial, repair history, fault code definitions, the monitoring timeline. Half of that list can be captured by the person at the pad in ten minutes, and the guide's own instruction is the one that matters: "Document everything from the moment there is any sign of a problem."
Picture the fleet engineer who opens a base-cabinet photo, spots the water line on the door, and cancels a crane standby that was booked on the alarm alone. The turbine still needs a visit, just not a lift.
I'd rather send a landowner to the base door with a six-step flow than book that crane on an alarm, and most O&M managers I talk to feel the same once they've paid for one idle crane day.
The asynchronous review loop
The reason this works on a wind farm at 06:40 is that nobody has to be on a call.
Your desk sends a secure, personal capture link by SMS, WhatsApp or email, or the QR sits on the base door and the kiosk. The site technician, the landowner or the contractor opens it in the phone's browser, follows the steps, and submits.
Your engineer reviews it at the desk when the case arrives. Asynchronous capture decouples the site's time from the engineer's time, which on distributed assets is the whole point. Send now, capture later, review when ready.
What the engineer gets is a structured case, not a folder of photos: every answer to every step, the transcript of what was read from the display, timestamps, device context, the asset reference from the scan, and a seal on the submission. From there the review and retake tools do the operational work:
- Retake in one click when the display shot is unreadable, with the site's own reason for re-recording stored on the case.
- Assign or escalate to the converter specialist, the blade engineer or the OEM contact, with an SLA in hours on the route if you want one.
- Copy a view link for the crane company or the OEM warranty desk, who need to see the case and have no login.
- Push it onward by API or webhook into the CMMS or ticketing system the work order lives in.
In support and field-service settings, teams using Venta Capture solve 23% of customer problems without dispatching a technician. A tripped turbine won't often be one of those.
What changes on the wind board is the quality of the dispatch: the right two people, the right module on the van, the crane booked on evidence rather than on an alarm, and the calm day used for the job that needs it. That is what moves first-time fix rate on assets where a second attempt costs a week.
Want to see it against one of your own fault types? Book a demo and bring a real alarm from last month.
Completion and generation confirmation
The same flow runs in the other direction once the crew has finished, and in renewables it earns its keep twice.
A completion capture turns "job closed" into a record that stands up to the O&M contract and the OEM warranty desk. After a converter swap or a combiner repair, the technician on site captures:
- The part: label and serial of the removed and fitted component, side by side.
- The work: torque marks, terminations, the cabinet closed and locked, the display back in run state with the time visible.
- Generation: the controller or inverter reading once the asset is back online, read out loud on video.
- The site as left: waste removed, access track and hardstanding as found, gate secured.
The mechanics are the same as any proof of work in field service, so they are not repeated here. What matters for an O&M manager is who reads the record afterwards.
The contract reads it first. kWh Analytics looked at 165 commercial and utility-scale solar systems and found mean availability between 96.4% and 97.2%, with some system-years below 80%. When availability and response times are contractual, a timestamped, sealed record of when the asset was found, when it was back online and what the reading was is the difference between an argument and an invoice.
The warranty desk reads it second. A warranty claim on a converter module or an inverter board goes better with the original display fault, the physical condition before the work, the serial numbers and the completion state all in one sealed submission with an audit trail.
The seal carries a SHA-256 fingerprint and a downloadable manifest, so the OEM can verify the file wasn't edited after the fact. The cross-sector rules for that kind of record are in the post on photo evidence standards for inspections.
What this does not do
Some limits, stated plainly, because O&M managers are rightly allergic to overclaiming.
- It does not diagnose. Guided capture gets better evidence to the qualified engineer. The fault call, the dispatch decision and the crane decision stay with that person.
- It does not replace SCADA or condition monitoring. It fills the gap between the alarm and the pad. Vibration, oil and electrical data keep doing what they do.
- It does not put anyone up a tower or inside an enclosure. The ground-level flow is exterior only. Up-tower steps are for certified technicians already there under their own permit.
- It does not remove visits. Remote-first, not remote-only. The aim is that the visits you do make are the right ones, with the right parts and the right lift.
- It does need mobile coverage. The flow runs in the phone's browser, so a pad with no signal at all is a pad you still visit blind. Where coverage is patchy, the person captures what they can from the nearest point that has it: the access track, the gate, the rotor from a distance.
- The timestamp proves receipt, not the fault. The server-verified time says when the case arrived, not when the trip happened. SCADA still owns that.
- It does not schedule itself. Periodic condition captures on remote assets work well, but your team sends the requests.
If that reads like a fair set of boundaries for your fleet, the fastest test is one flow on one fault type. Start for free, build the ground-level turbine flow above, and send it against the next trip.
Rooftop and commercial PV is covered in solar inverter fault remote diagnosis; the cost logic behind avoided site visits in the cost of an unnecessary field service visit.
Frequently asked questions
Does guided capture replace SCADA or condition monitoring?
No. SCADA and condition monitoring report the asset's electrical and mechanical state from a distance, and guided capture shows your engineer the physical state of the pad, the cabinet and the display, which the data line cannot carry. You need both to decide a dispatch well.
Can a landowner or a security contractor really capture something useful?
Yes, as long as the flow only asks for what they can safely see from the ground: the base door, the display through the window, the track, the visible rotor, the sound. The expertise sits in the steps you build, not in the person holding the phone. That is the whole idea behind using the person who is already there.
What about sites with no mobile signal?
The flow runs in the phone's browser, so it needs coverage to work. On a pad with none, capture what you can from the nearest point with signal (track, gate, rotor from a distance) and treat the rest as a visit. The server-verified receipt time is shown on every case, so nobody has to guess when it came in.
Is anyone asked to climb or open an enclosure?
No. The ground-level flow is exterior only and says so on its first step. Up-tower and inside-enclosure steps exist only for certified technicians who are already there under their own procedure, and the flow is a checklist for them rather than an instruction to anyone else.
How does the completion record help with an OEM warranty claim?
The claim carries the display fault as found, the physical condition before the work, the removed and fitted serials, and the asset back in run state, all in one sealed case with a fingerprint the OEM can verify. Warranty desks argue less with a record they can check than with a folder of photos and an email.
What happens if the capture is not enough to decide?
The engineer requests a retake with a specific instruction, or the visit goes ahead with the evidence it has. Avoiding every visit was never the goal; sending the right people with the right parts, and booking a crane only when the case shows one is needed, is.
How long does it take to set up a flow for one fault type?
A standard flow with a scan step, a few photo slots with example images and a video step can be live the same day; the free plan needs no credit card and the first standard flow can be built in a free setup call. Flows with branching logic, routing rules or integrations into your CMMS are set up around your process by VentaVid as a paid project.
Talk to us
If the next alarm on your board will cost a drive, a climb or a crane before anyone knows what it is, test the other order on one fault type. Build the ground-level flow, send it to whoever is nearest on the next trip, and see what your engineer can decide before the van leaves.
What it takes to start.
- A free plan, no credit card, live in about 10 minutes.
- Nothing for the site technician, landowner or contractor to install. It runs in the browser on the phone they already carry.
- Stuck? Book a free setup call and we build your first standard flow together.
Venta Capture is built by the VentaVid team, who have spent more than a decade putting video into sales and service workflows for teams in 43 countries.
Start for free and build your first turbine or inverter flow, or book a demo and we'll map it to your dispatch process. The wider field service use case and the request-by-link mechanics are on the product pages.