← All articles
Federal / VA

Fragnets and TIA on federal projects: the right way to request a time extension

On federal work a time extension is earned through the network, not argued in a letter. You build a fragnet for the change or delay, insert it into the current accepted update, and let the time impact analysis show what it does to the critical path. The sequence, and the mistakes that get requests denied.

CPM Predictive Solutions·7/2/26·7 min read

Federal schedule specs, including the UFGS and VA NAS sections, require time extension requests to be supported by a time impact analysis: a fragnet representing the change or delay, inserted into the update in effect when the event occurred, with a before-and-after comparison of the completion date. If the event drives the critical path and moves completion, you have the basis for a time request, subject to responsibility, concurrency, notice and the contract. If it only absorbs float, you don't, however disruptive it felt. The Contracting Officer grants days the network supports, not days a narrative asks for.

What a fragnet is

A fragnet is a small piece of network, the activities and logic for the changed or delayed work, built like the rest of your schedule. For a design change it is the new work (submittals, procurement, added installation) tied into existing logic where the work connects. For a government-caused delay it might be a suspension activity or a differing site condition investigation, tied in where it interrupts the flow. An unsupported delay bar sized to the requested extension assumes the answer. The fragnet's detail and durations must match the event and the record.

The TIA sequence, step by step

The prospective method, which federal specs call for and AACE RP 52R-06 addresses:

  • Start from the right schedule. The most recent accepted update with a data date immediately before the event. Not the baseline from a year ago, and not a schedule you adjusted for the occasion. The unimpacted run of that update is your "before."
  • Build the fragnet. Activities and logic for the change or delay, with durations you can support from quotes, submittal timelines or crew production.
  • Insert and recalculate. Tie the fragnet into the network and run it. The new completion date against the old one is the impact, measured in CDs.
  • Read the result honestly. If completion moved, the difference is your requested extension. If completion held but total float on the affected paths dropped, the event consumed float. Under most federal specs that isn't compensable time, but it belongs on the record for the cumulative picture.
  • Package it. The fragnet, before-and-after runs, a narrative connecting the event to the government action or change, and coding that ties it to the modification or REA. Submit inside your spec's request window.
TIME IMPACT ANALYSIS · REA-004PROSPECTIVE · AACE RP 52R-06
RUN A · ACCEPTED UPDATE · DD 01 JUN · UNIMPACTED DATA DATE DUCT RISER L3 AHU SET + CONNECT TEST + BALANCE COMPLETION 14 NOV driving path TF 0 · saved as the BEFORE condition RUN B · SAME UPDATE + FRAGNET REA-004 INSERTED DUCT RISER L3 FRAGNET REA-004 · DESIGN REVISION RFI/REV 12d PROCURE 14d INSTALL 9d TEST + BAL +18 CDs RUN COMPLETION DRIVING PATH TF DELTA A · unimpacted 14 NOV 0 CDs base B · fragnet in 02 DEC -18 CDs +18 CDs modeled impact
The package the reviewer wants: Run A saved as the before condition, the multi-activity fragnet threaded into the logic, and the run table showing the modeled effect, backed by the narrative and records.

Float ownership: the rule that surprises contractors

Federal specs almost universally treat total float as a shared project resource, belonging to neither the contractor nor the government. An event earns a time extension only when it hits the critical path and moves completion. A government change that eats 15 CDs of total float on a near-critical path earns nothing today, and the government can consume float the same way you can. So document every impact, compensable or not, when it happens. Float consumption compounds, and the change that finally goes critical is far easier to support when the record shows the path being eaten one modification at a time. We covered that trap from the private-work side in the change order release article, and the mechanics are identical.

The mistakes that get requests denied

  • Wrong base schedule. Inserting the fragnet into the baseline or a stale update instead of the accepted update contemporaneous with the event. Measure against where the job actually stood.
  • The one-bar fragnet. One activity whose duration equals the days requested. Reviewers see through it.
  • Unsupported durations. Fragnet durations pulled from the air. Tie them to submittal registers, vendor quotes, or production rates.
  • Late submission. Specs set a window for time extension requests, and on VA and USACE work the request is expected while the impact is prospective, not reconstructed at the end. Notice first, analysis promptly after. The notice principles apply in full.
  • Ignoring concurrency. If your own delay runs alongside the government's, the analysis has to address it. Ignoring it hands the government its denial.
  • No tie to the mod. The fragnet and request should carry coding that ties them to the specific modification or REA, so time and cost move together through the government's process.

Field notes from the federal cycle

  • Get the fragnet approved before it enters the update. Many federal specs require government approval of fragnets before they go into the schedule of record. Submit the fragnet with the TIA as its own package, get concurrence, then incorporate. Merge an unapproved fragnet and a clean analysis becomes a rejected update, and both get stuck.
  • Give fragnet activities their own ID band and code them to the mod. A dedicated ID prefix per REA or modification keeps the fragnet auditable years later and your SDEF export clean, and lets you filter a change's full history in seconds when negotiation comes.
  • Use the update just before the event. For a prospective TIA, status the schedule to the day before the event, save, then insert the fragnet. For a retrospective window, the half-step (progress first, then the event) separates delay from your own slippage, and it is the first thing a forensic analyst tests.
  • Net out anticipated weather before you ask. Federal contracts carry anticipated adverse weather, often as a monthly day table. A TIA claiming gross weather days without netting the allowance gets the whole request discounted. Show the math: actual less anticipated, then the effect on controlling work.
  • Read the modification like a change order, because it is one. Bilateral mods carry release language just like private CO forms. Signing for the direct cost with silence on time can close the door the TIA just opened. Reserve time and impact rights on the mod itself, in writing, every time.

Where this sits in the federal chain

On agency work the TIA is part of the rhythm the spec sets: the accepted cost-loaded baseline, the monthly update cycle, and fragnet-based TIAs whenever changes or government delays hit, each coded to its REA or modification. Run that rhythm and a time extension is usually a routine contract action. Skip it and the same request tends to become a dispute. It's the standing scope we carry on federal projects, and the same discipline drives our change order and TIA service on private jobs. REAs and certified claims sit under the FAR and the disputes process, counsel advises on framing, and certification is signed by a person authorized to bind the contractor on the claim. The analysis underneath, the part the CO's schedule reviewer will test, is ours.

Change or delay hit your federal job?

Your spec's request window is running. Send the mod or event description and the current accepted update, and we'll build the fragnet and run the impact before it closes.

Frequently asked questions

What is the difference between a fragnet and a TIA?
The fragnet is the piece of network representing the change or delay. The time impact analysis inserts it into the current accepted update and measures the effect on completion. The fragnet is the input, the TIA is the analysis and result.
Which schedule do I insert the fragnet into?
The most recent accepted update with a data date immediately prior to the event. Measuring against the baseline or a stale update misstates the impact, and reviewers check this first.
The change ate our float but did not move completion. Do we get time?
Under most federal specs, no. Total float belongs to the project, and an extension is earned when the critical path moves. Document the float consumption anyway, in the TIA and the update narrative, because that cumulative record supports the request when a later event finally drives the path critical.
Can we wait and prove the delay at the end of the project instead?
Federal specs expect prospective, contemporaneous requests. Waiting usually means a forensic fight later under AACE RP 29R-03 methods, with a weaker position on notice and timeliness. The contemporaneous TIA is cheaper, faster and far more likely to be handled as a routine contract action.

Need a fragnet built and an impact run this week?

Send the modification or delay event and your current accepted update. On a 20 minute call we'll scope the fragnet, the records it needs and the turnaround.

Back to top ↑