the wind agent.

Reading the wind.

Checking your session…

HOW THE FLEET BOARD WORKS

One screen for every site, sorted by the next bad hour

The fleet board is for anyone who answers for more than one place: a crane hire yard with machines on six sites, a contractor with three pours and a tower, a spraying firm, a marine operator, an events team. Each saved site becomes one row. Each row carries the site's working height, its limit, a lamp, and the first hour the odds of crossing that limit reach 40%. The board is sorted so the site that will need a decision soonest is at the top.

What is on a row

  • Now. The modelled wind at the site's working height for the current hour, in your unit, with the method printed: native model level, interpolated or extrapolated.
  • Limit. The threshold from the site's rule, with its metric (mean speed or gust) and comparison (at or above, or at or below).
  • Next crossing. The first hour in the shift where the ensemble odds reach 40%, or “No crossing in shift”.
  • Peak odds. The highest ensemble fraction in the shift, with the member count and the ensemble height it came from.
  • Lamp and reason. GO, WATCH, STOP or UNKNOWN, with one line saying why.

The lamp logic, exactly

The lamp is computed in the browser from two things only: the deterministic forecast and the ensemble. There is no hidden score. For each hour of the shift it takes the highest ensemble fraction of members that break your rule, and checks whether the primary forecast breaks it.

LampWhen it shows
STOPThe primary forecast reaches the limit in any shift hour, or the ensemble odds reach 70% or more in any shift hour.
WATCHEnsemble odds reach 40% in any shift hour, or the primary forecast comes within 5 mph of the limit without crossing it.
GONeither of the above across the whole shift: the forecast stays inside the limit and the odds stay under 40%.
UNKNOWNNo rule set, forecast unavailable, ensemble unavailable so the odds were not evaluated, or missing hours in the shift. The row says which.

Rules that set a floor (“at or below”) are handled the other way up: the odds used are the fraction of members that fall under the threshold. The odds are raw, uncalibrated ensemble fractions, so the lamp is a prompt to look at the instrument for that site, not a verdict.

Shift windows

Weather risk matters during the hours you are working. The shift selector sets a start and an end hour in site-local time; the default is 06:00 to 18:00. Only forecast hours inside that window are evaluated, starting from the current hour. Overnight shifts, such as 22:00 to 06:00, are supported and are treated as one contiguous run across midnight. A site in Dublin and a site in Calgary therefore each use their own wall clock.

Rules: what a site is measured against

A rule has five parts, set from the pencil on any row or from Account.

  • Metric: mean wind speed or gust.
  • Comparison: at or above the threshold (most limits) or at or below it (minimum wind for something like a wind-driven process).
  • Threshold: the number, entered in your unit and stored in mph.
  • Height: the working height in metres above ground, for example 45 m for a jib or 100 m for a drone.
  • Source document: where the limit comes from, such as a lift plan, a manufacturer's manual or a method statement. The board keeps the name next to the number, so anyone reading it sees whose limit it is.

The Wind Agent holds no limits of its own. It will not suggest a threshold or tell you that a crane may work at a given speed. The number is yours, from your document.

A worked day

A crane firm saves four sites with a 27 mph gust limit from the manufacturer's manual, and sets the shift to 07:00 to 17:00. At 06:30 the board shows Site A as GO: gusts peak at 19 mph and the highest odds are 4%. Site B is WATCH: the primary forecast is within five miles per hour of the limit at 14:00. Site C is STOP: 11 of 15 hours are fine but the ensemble puts 74% on a gust of 27 mph or more at 15:00. Site D is UNKNOWN because the ensemble feed did not answer and the row says so. The supervisor opens Site C in the instrument, sees a wide ensemble and a SPLIT chip, brings the lift forward to the morning and records the decision. The board did not make that call. It put the right site at the top.

What each plan includes

PlanSaved sitesForecastAlerts and briefFleet boardAPI / day
starter37 daysNoNonone
pro2516 daysYesNo1,000
ops10016 daysYesYes10,000
enterprise50016 daysYesYes100,000

Related reading: setting a limit from a document, height-matched wind for cranes and ensembles and probability.

Fleet board questions

What does the fleet board show that the instrument does not?

The instrument looks at one place at a time. The fleet board holds every saved site on one screen, each at its own working height against its own rule, and sorts them by the next hour the odds cross 40%. It answers which site needs attention first.

What do GO, WATCH and STOP actually mean?

STOP means the primary forecast reaches the limit during the shift, or the ensemble odds reach 70% in some shift hour. WATCH means the ensemble odds reach 40%, or the primary forecast comes within 5 mph of the limit. GO means neither happened across the whole shift. UNKNOWN means the board could not evaluate the site and says why. None of them is a clearance.

Why is a site showing UNKNOWN?

Either it has no rule yet, the forecast did not load, the ensemble was unavailable so the odds were not evaluated, or hours in the shift were missing data. The reason is printed on the row. Missing data never counts as calm.

Does the board use the working height of each site?

Yes. The speed shown is the modelled wind at the site's working height, interpolated between the 10, 80, 120 and 180 m model levels. Ensemble odds use the nearest ensemble level (10, 80, 100 or 120 m) and the board shows which level was used. Gust odds are always at 10 m, because the models define gust there only.

Which plans include the fleet board?

Ops and Enterprise. Ops holds 100 saved sites and 10,000 API calls a day; Enterprise holds 500 sites. Starter and Pro have the instrument, records and, on Pro, email alerts and the morning brief.

Is the board a go or no-go decision?

No. The lamp compares a modelled forecast with the limit you typed in from your own document. Your threshold, your procedure and your call: the board prepares the decision and does not make it.