Ice storms are won or lost before the first drop freezes. Watch the 72 hours in front of a major icing event, 0.7 inches of radial ice with 30 mph wind, where every decision has an expiration time: the mutual assistance request that gets 200 crews only if it beats the neighboring utilities to the phone, materials pushed forward before roads glaze, transmission anti-icing by re-routing load current through at-risk lines, a heating-surge load forecast that says +38%, and a rotating-outage plan you build and hope to never touch. Then the storm arrives, and the utility that spent its 72 hours wisely restores in 2.3 days instead of 6.5.
| Without | With GridCORTEX | Δ |
|---|
| Without | With GridCORTEX | Δ |
|---|
A fictional utility's winter weather desk receives an outlook 72 hours ahead of a major ice storm: freezing rain is likely, expected to build 0.5 to 0.7 inches of ice around every wire across the northwest band, with a 30 mph north wind behind it, at 68% confidence and rising. Ice that thick snaps wires and poles, and the wind then whips coated lines until they fail. GridCORTEX turns the forecast into a damage estimate: roughly 1,900 broken spans of line (a span is the stretch of wire between two poles), concentrated where heavy ice, overhanging trees, and older wire stack on top of each other. The simulation clock runs 120 hours: the first 72 count down to the ice, labeled T-72 to T-0, and the rest play out the storm and the repair effort through 48 hours after the ice arrives. The premise of the whole demo: ice storms are won or lost before the first drop freezes, because every preparation move has a deadline.
Four decision points follow the countdown, and a person, the storm director, approves each one, because each one commits real money and real crews before the storm is certain. At T-66, 66 hours before the ice, the first call: five utilities sit in this storm's path, and mutual assistance, the industry system where utilities lend each other repair crews, goes to whoever asks first. The recommendation is to request 200 borrowed crews now, move poles, wire, splices, and transformers out to the Northgate, Midtown, and Riverside staging sites before the roads ice over, and book crew lodging. Confirmation comes within hours: 200 crews from three states, while a neighboring utility that called later gets 40 and a waitlist. At T-40, confidence is 88% and the forecast says demand will jump 38% as electric heating works harder in the cold. The second approval starts anti-icing: extra electric current is routed through the three most at-risk high-voltage lines so the wires warm themselves 2 to 3 degrees Celsius above the temperature at which ice can form. The utility also commits all available generation, verifies fuel, and pre-drafts a public appeal to conserve power. At T-12, the last window when the roads are still safe, the third approval moves crews into shelter at the staging sites, checks the backup generators at hospitals and water plants, and builds the rotating-outage plan nobody wants to use: 29 groups of customers who could be switched off in turn for short periods if demand outruns supply, computed in advance, with water plants excluded.
The freezing rain begins on schedule at T-0. Six hours in, ice is at 0.3 inches, 9,800 customers are out, and two high-voltage lines are alarming because their ice-coated wires have started bouncing in the wind, a failure mode called galloping. Ten hours in, the fourth decision point launches the repair sequence. Lines are restored in order of criticality, and carefully: after a long cold outage, everything in every home switches back on at once, so each re-energized line briefly draws about 1.6 times its normal load, and pickups are spaced 90 seconds apart to protect transformers. The borrowed crews are already at their pre-assigned sectors. And drones fly de-icing passes over the Riverside-Bayshore corridor, the one line that could not be warmed without overloading it, knocking the glaze off span by span and confirming each pass with laser scanning before the evening crosswind. The event peaks 16 hours in: 0.7 inches of ice, 31,000 customers out, demand at 97% of available supply, and the rotating-outage plan stays parked.
With all four approvals, 18,000 customers are back 26 hours after the ice starts, and by hour 40 the system is 92% restored with zero lines failing again under the surge of returning load. Full restoration is projected at 55 hours, 2.3 days, and the after-action report, every pre-storm decision timestamped with the forecast confidence it was made on, is already assembled. Decline the recommendations and the same storm runs 6.5 days, with 78,000 customers out at the peak and 2 hours of improvised rotating blackouts at evening peak.
The conventional storm room waits for certainty: "let's see the 48-hour forecast first." Every call in the playbook gets made, but each one after its deadline has passed. The crew request goes out on storm day, lands in a queue behind four other utilities, and returns 38 crews arriving in 2 to 3 days. Materials sit in the central yard behind iced roads. The at-risk high-voltage lines run cold and lightly loaded, so they glaze over like everything else.
When the 0.7-inch band lands, two high-voltage transmission lines break under the weight of the ice, blacking out whole substations: 78,000 customers out, demand at 99% of supply, and 2 hours of rotating blackouts improvised from a spreadsheet at evening peak. Restoration crawls on borrowed crews and sliding convoys. Two lines fail again under the surge of returning load and have to be switched back off. At hour 46 the system is only 61% restored, the projection says 6.5 days, and the state regulator's letter arrives before the last customer's power does.
The software converts icing physics into decisions with deadlines attached. It models the damage span by span to size the response, and lands within 8% of the actual break count. Its confidence thresholds trigger the T-66 request that wins 200 crews. Routing current through the three at-risk lines keeps them warm and turns would-be casualties into survivors. Its restoration sequencing brings customers back in an order that never overloads a transformer. Every play needs a person: the storm director approves all four recommendations, and the emergency operations team executes them.
The winning numbers: 0 transmission lines lost, 31,000 customers out at peak instead of 78,000, roughly 1.4 million customer-hours of outage instead of 5.9 million (one customer-hour is one customer without power for one hour), zero repeat failures, a blackout plan computed and never used, and full restoration in 2.3 days instead of 6.5.
| Measure | Without GridCORTEX | With GridCORTEX | The difference |
|---|---|---|---|
| MA request timingMA is mutual assistance, the system where utilities lend each other repair crews; asking first decides who gets them | storm day: 38 crews | T−66: 200 crews | the clock won |
| Materialsthe poles, wire, splices, and transformers repairs need, and where they sat when the roads iced over | central yard, iced roads | staged forward, dry roads | pre-positioned |
| Transmission corridorsthe high-voltage backbone lines; losing one blacks out whole substations at once | 2 down under ice | kept warm by routed current, 0 down | current used as heat |
| Crew posturewhere the repair crews were sheltered when the storm hit | driving on ice at 3 AM | sheltered at the staging sites | safety + speed |
| Drone de-icingdrones knocking ice off the one line that could not be warmed, each pass confirmed by laser scan (LiDAR) | no such capability | 5 spans, LiDAR-verified | see The Fleet Above |
| Shed planthe rotating-outage plan: switching customer groups off in turn for short periods when demand outruns supply | improvised at 99% load | computed, parked, unused | readiness |
| Damage model accuracyhow close the pre-storm estimate of broken spans came to the real count | n/a | within 8% of actuals | staging was right |
| Full restorationthe time until the last customer's power is back | 6.5 days | 2.3 days | 4.2 fewer days dark |
| Peak customers outthe most customers without power at any one moment | 78,000 | 31,000 | the backbone held |
| Customer-hoursthe total outage burden: one customer without power for one hour counts as one customer-hour | ~5.9M | ~1.4M | 76% less time in the dark |
| Cold-load re-tripslines that fail again under the surge of everything switching back on at once after a cold outage | 2 lines had to be switched off again | 0; pickups carefully spaced | engineering |
| Rotating outagesplanned short blackouts rotated between customer groups when supply cannot cover demand | 2 hrs at evening peak | none | the parked plan |
| After-action packetthe report to the regulator showing every storm decision and when it was made | 3 weeks to assemble, contested | pre-storm decisions traced | defensible |
The safety effect here is indirect and worth saying plainly. Nobody is removed from a hazard by a dashboard. What changes is that crews get dispatched against a confirmed picture rather than a stale one, so fewer crews drive to a location that was already restored or that turns out to be a different job than the one they were sent for.
Counted in units you already track:
Status collection and reporting hours come back to the people running the event, and the EOC director gets the projection built for them instead of assembled by them.
The numbers we need from you to run that formula:
| Cost driver | How it is calculated, from a rate you supply |
|---|---|
| EOC staffing | collection and reporting hours avoided x your loaded rate for those positions x event hours per year |
| Crew productive time | crew hours recovered from misdirected moves x crew size x your loaded crew rate |
| Restoration duration | hours cut off the event x your fully loaded restoration cost per hour, including contractor and mutual assistance crews already on the clock |
| Mutual assistance right sizing | crew days released earlier x your all in mutual assistance cost per crew day. You decide how many days a better picture actually releases |
You pay for the scoped engagement that builds and runs this, for integrations into SCADA, the outage management system, crew scheduling, weather, and whatever imagery feed you use, and for your own staff time to run it alongside the current process through at least one real event. The number of integrations is the cost driver, and crew status is usually the hardest one, because it is the least standardized system you own.
Payback is driven by hours cut off event duration and by EOC staffing hours, in that order. Event duration is the larger number but the harder one to attribute, so build the base case on staffing hours and treat duration as upside you validate over a season.
Worker exposure is not the main story here and we will not pretend otherwise. The direct safety mechanism is public: life support customers, water pumping, and emergency communications stay energized because the exclusions are enforced by the plan rather than by an operator's recall under extreme pressure, and no block gets held past its planned duration because a timer was missed.
Counted in units you already track:
Plan building and block tracking hours come back to the operators and support staff working the emergency, and reporting hours come back to the regulatory team afterward.
The numbers we need from you to run that formula:
| Cost driver | How it is calculated, from a rate you supply |
|---|---|
| Emergency staffing | activation hours avoided x staff involved x your loaded rate at the overtime and callout premiums that apply during an emergency |
| Regulatory reporting | post-event reporting and data request hours avoided x your loaded regulatory analyst and legal support rate |
| Customer care volume | calls avoided through accurate, block-specific notification x your own fully loaded cost per contact center call |
| Over-shed energy | megawatt hours shed beyond the required amount x your own value of lost load or your own cost per unserved megawatt hour, a figure you set |
| Switching device duty | rotation operations avoided x your maintenance cost per operation, since operation counts drive recloser and breaker inspection intervals |
You pay for the scoped engagement that builds and runs this, for integrations to your ADMS, advanced distribution management system, your customer information system, and your GIS, and for the designation work underneath: somebody on your side has to establish and defend which feeders are critical and which customers are medical baseline, and keep that current. Add drill time, because a tool nobody has practiced with is not going to be trusted during the one hour it matters.
Payback here is awkward and you should hear it straight: capacity emergencies are infrequent, so an hours-per-event case built on frequency will not hold up. Build it on the preparation and the post-event regulatory reporting work, which happens whether or not you shed, and treat the emergency day savings as the reason it exists rather than the reason it pays.
A crew that has driven several hundred miles and then waits half a day is a fatigued crew starting late, and a crew released to work before it has absorbed your local rules is a crew operating on another utility's practices in your territory. Compressing orientation and getting assignments right at arrival is a safety control, not just a logistics improvement.
Counted in units you already track:
The largest block of hours returned is not yours, it is the arriving crews' idle time between arrival and first assignment, and you are paying for those hours at a contracted rate.
The numbers we need from you to run that formula:
| Cost driver | How it is calculated, from a rate you supply |
|---|---|
| Idle crew time | wait hours removed x personnel x the contracted rate you pay assisting crews, which you pay whether they are working or waiting |
| Coordinator labor | coordination hours avoided x your loaded coordinator rate, multiplied by the number of coordinators an activation consumes |
| Reconciliation and accounting | post event reconciliation hours avoided x your loaded accounting rate |
| Cost recovery exposure | your own history of disallowed or delayed storm cost recovery x the share you attribute to incomplete documentation, a number your regulatory accounting group already knows |
| Lodging and per diem | crew days removed from the event x your lodging and per diem cost per person per day |
You pay for the GridCORTEX coordination assistant, for integration into your OMS (outage management system) work queue, your cost and timekeeping systems, and your email, and for your coordinators to keep the orientation content current, since your safety rules, radio channels, and lodging arrangements change between events. The content upkeep is small but it is not zero, and stale packets are worse than none.
Payback is dominated by idle crew hours at the contracted assisting rate, because that is the largest and best documented number in an activation. Coordinator labor and reconciliation are real but secondary, and cost recovery exposure should be sized by your regulatory accounting group, not by us.
The safety effect is indirect and it comes from converting emergency winter work into planned fall work. Repairs made during a declared emergency at design minimum temperatures, on ice, at night, are the most dangerous work a plant or line organization does, and every freeze failure prevented is a set of those jobs that never gets written.
Counted in units you already track:
Scenario setup and analysis hours come back to the generation planning group, and the resilience planner receives a ranked action list instead of building the study that produces one.
The numbers we need from you to run that formula:
| Cost driver | How it is calculated, from a rate you supply |
|---|---|
| Replacement power and scarcity exposure | megawatts of shortfall avoided x hours of the event x your own scarcity or replacement power price. In a real cold event this dwarfs everything else and it is entirely driven by your market position |
| Analyst labor | scenario setup and run hours avoided x your loaded analyst rate |
| Emergency versus planned work | emergent winter crew hours converted into planned fall hours x the difference between your emergency and straight time loaded crew rates, including callout and overtime premiums |
| Winterization spend targeting | units or components not winterized because the model shows no exposure x your per unit winterization cost, offset by the units the model adds to the scope |
| Load shed exposure | customer minutes of firm load shed avoided x your value per customer minute, if you serve load. Only you can decide how much of an avoided shed to attribute to preparation |
You pay for the simulation service and the compute behind it, for integration into your generation, fuel, transmission, and weather data, and for your own planners' and plant engineers' time validating that the modeled failure modes match what your units actually do. Getting unit level cold weather performance data into usable shape is usually the first season's real work.
Payback is dominated by a single avoided shortfall event, which means this is really an insurance case and should be presented that way. The scenario analysis labor savings are steady and small, and the shortfall avoidance is lumpy and large.
The two highest consequence exposures in line inspection are low altitude manned flight and climbing energized structures. Both are replaced for routine condition assessment, and the climber only goes up once the defect is already confirmed and the repair is scoped.
Counted in units you already track:
Patrol hours and desk triage hours come back to the line department, and the inspection engineer stops reading every image and starts reading only the flagged ones.
The numbers we need from you to run that formula:
| Cost driver | How it is calculated, from a rate you supply |
|---|---|
| Patrol labor | patrol hours avoided x your loaded patroller rate |
| Aviation | manned inspection hours avoided x your all-in helicopter hourly cost, whether owned or contracted |
| Engineering triage | triage hours avoided x your loaded engineer rate |
| Avoided failure | your own cost per unplanned transmission outage x the share of failures you believe earlier detection would have caught, which you set, not us |
| Vehicle and travel | road miles avoided x your fleet cost per mile, plus per diem on remote patrols |
You pay for the flight operations, whether you own the fleet or contract it, for the GridCORTEX intelligence layer that turns imagery into ranked findings, for the integration into your asset and work systems, and for your own staff time to validate detections during the first inspection season. The intelligence layer is the smaller line item. The flight operations dominate.
Payback is usually driven by aviation and patrol labor, not by avoided failures, because avoided failure is the number you will trust least. Build the case on the two you can audit and treat avoided failure as upside.
What is this, exactly? It is AI software: intelligent agents and models built and delivered by SoftServe, running on NVIDIA accelerated computing. It is not a hardware appliance and it does not replace the systems you run today. It deploys in your own cloud or on your premises, connects read-only to your existing systems, and recommends; your people approve every action, starting in shadow mode until it earns trust.
A simulation service for operations and resilience planners that models your grid under extreme cold and returns a ranked list of preparation actions, with the failures each action avoids. The demo above uses synthetic data; everything below describes what the real deployment needs from your organization.
| Your system | Typical products | How we connect |
|---|---|---|
| Energy Management System (EMS) / transmission SCADA | GE e-terra, AspenTech OSI monarch, Hitachi Energy | scheduled file export (CSV or CIM XML) |
| SCADA historian | AVEVA PI System, AspenTech eDNA | historian mirror (one-way feed) |
| Planning and study tools | PSS/E, PowerWorld, TARA | scheduled file export (CSV or CIM XML) |
| Weather and environment | National Weather Service feeds, commercial forecast services | read-only API |
| Asset / work management (EAM/CMMS) | IBM Maximo, SAP PM | database replica refreshed nightly |
| Outage Management System (OMS) | GE PowerOn, Oracle NMS, ADMS outage module | event stream (read-only) |
| Field and crew systems | ARCOS, mobile workforce tools, vehicle location (AVL) | read-only API |
Runs in your own cloud account on GPU instances, or on an on-premises NVIDIA server. All connections are read-only through your existing data zone, with no connection to control systems and no control actions; it starts in shadow mode, replaying past winters.
The Approve button you just clicked in the demo above is the real workflow. This is what it looks like on the screen of the operations resilience planner in the GridCORTEX console:
Approve creates draft winterization work orders in your asset management system through its API, in pending status. Your maintenance planner reviews, schedules, and releases them under normal work management rules; GridCORTEX never releases work itself.
Scenario runs trigger automatically from the connected weather and historian feeds; to test a custom scenario, enter a temperature and duration in the console.
Reads EMS telemetry and the historian continuously and weather every 15 minutes; each card shows the as-of timestamp of its data. Early shadow pilots may run on periodic replicas, with the cadence shown on the card.
Scenario runs live in the GridCORTEX console linked to planning tools; a modeled shortfall above threshold sends an email and Teams push to operations leadership. The console runs in a browser beside your existing screens on day one; embedding into your own systems is a roadmap step once the read-only phase has earned trust. Approve, Modify, and Decline are all captured in an audit trail your compliance team can pull, and GridCORTEX never blocks or overrides anything in the systems you run today.
The fair question from any storm director: "We have weather vendors, an EOC playbook, and mutual assistance agreements, what's new here?" Here's the honest answer.
When someone asks "what did it actually calculate?", this is the list. In the simulation these factors drive the storyline; in a pilot they are computed from your territory, climatology, and agreements.
Presenter's one-liner: "The forecast said seven-tenths of an inch with wind. The twin turned that into a damage map, the damage map into two hundred mutual assistance crews requested a day before anyone else, materials on the right side of the ice, three transmission lines kept warm with routed load, and a rotating-outage plan we never had to touch. The ice came on schedule. We were done in two days; the utility next door took a week."