Skip to main content

Leak-Repair Episodes

A leak-repair episode is TuffOps' record of one cycle of finding a leak, fixing it, and proving the fix held. The lifecycle is driven by EPA timelines, not by user action — the system opens, advances, and warns about deadlines automatically.

This page explains the lifecycle so you know what the system is doing on your behalf.

How an episode opens​

Every refrigerant addition recomputes the unit's leak rate over a rolling 12-month window. When the calc lands at or above the §84.106(d) threshold for the unit's equipment category, the leak-rate observer:

  1. Stamps the calc as red.
  2. Opens a new regulatory_leak_repair row, snapshotting the discovery state — the leak rate, the threshold, the unit's full charge, and the calc that triggered it.
  3. Stamps the §84.106(d) repair due date based on the episode's deadline basis.
  4. Surfaces the new episode in the Leak Repairs index and on the unit's compliance panel.

There is no manual "open episode" button. Episodes always trace back to a calc.

The deadline basis​

The repair window depends on which subsection applies:

  • §84.106(d) — 30 days from discovery. This is the default basis for new Part 84 episodes.
  • §84.106(d) industrial-shutdown — 120 days from discovery. Applies when the unit is industrial-process refrigeration and the discovery context was an authorized industrial shutdown.
  • §82.157(c)(1) — 30 days from discovery, for the parallel Part 82 path. Applies for ODS-only carve-out units that still trip the leak rate, and for units where Subpart C applicability is not yet known.

The episode's deadline_basis column records which one applies, and deadline_window_days carries the resolved number of days (30 or 120).

Verification has two stages​

After the repair is logged, the episode owes two verifications:

  • Initial verification — under §84.106(e)(1), due on the same window as the repair clock (30 days from discovery, or 120 days under industrial shutdown). Confirms the repair held under operating pressure.
  • Follow-up verification — within 10 days of the successful initial verification under §84.106(e)(2). Confirms the system stayed sealed.

Each stage records the date, method, result, and notes. A failed initial verification keeps the episode open and resets the follow-up clock; the supervisor schedules a re-repair.

Three close paths​

An episode can close under one of three paths, recorded as the close_reason:

  • repair_verified — the standard happy path: repair logged, both verifications passed.
  • Retrofit / retire / mothball — under §84.106(h), the operator can commit to retrofitting, retiring, or mothballing the unit instead of repairing. The plan card on the episode page records the commitment date and computes the 1-year completion deadline (per §82.157). When plan_completed_at is set and the unit is updated to reflect the new state, the episode closes.
  • Other terminal reasons — cancelled, out_of_scope, etc., for episodes that opened on bad data or units that left the operator's scope.

Once closed, episodes cannot be reopened. A subsequent red calc opens a fresh episode.

The next-add presumption​

§84.106(d)(2) gives operators a presumption of compliance after a verified repair: if the next refrigerant addition lands the leak rate below the threshold, the repair is presumed to have worked.

TuffOps stamps this automatically. After an episode closes with both verifications passed:

  1. The unit waits for its next refrigerant addition.
  2. The leak-rate calc on that addition is checked.
  3. If the calc is not red, the Next-add presumption card on the closed episode stamps presumption_at and links to the confirming calc.
  4. If the calc is red, the presumption is rebutted — presumption_rebutted_at and the rebutting calc are stamped — and a new repair episode opens automatically.

The presumption card is read-only. The system manages it; you read it.

Extensions and shutdowns​

Two mechanisms can extend a §84.106(d) deadline:

  • §84.106(f) extension request — when parts are on order, the site is unreachable, or another documented cause makes the deadline impossible. Filed by a supervisor with compliance.manage_leak_repair_extension. The extension card on the episode page records the request, lets the operator withdraw it, and produces an EPA-formatted notification payload for the §84.106(f)(2) filing.
  • Industrial shutdown — when the unit is offline because the wider process is shut down. Filed by a supervisor with compliance.manage_leak_repair. Pauses the deadline for the shutdown duration.

Both mechanisms are visible on the episode page and recorded in EPA exports.

What the supervisor never has to do​

The system handles, with no user action:

  • Opening the episode when a calc goes red.
  • Stamping the §84.106(d) repair due date.
  • Stamping the §84.106(e)(1) initial verification due date when the repair completes.
  • Stamping the §84.106(e)(2) follow-up due date when the initial verification passes.
  • Computing the §82.157 1-year completion deadline when a retrofit/retire plan is committed.
  • Stamping or rebutting the §84.106(d)(2) next-add presumption.
  • Opening a new episode when the presumption is rebutted.

The supervisor logs the work and decides the path. The dates take care of themselves.