Methodology, Scope, and Validation

This page describes what the calculator does, where it stops, and what evidence supports a result. It is written for reviewers who need more than a number on a screen. The product is calculation and document support for a qualified-person workflow; it is not a professional-engineer seal, an energized-work authorization, an equipment-condition assessment, or a substitute for the employer's electrical safety program.

Last reviewed: July 19, 2026. Engine behavior is versioned separately from editorial content. A stored calculation record identifies the engine version used for that run.

Method selection

For systems inside the implemented IEEE 1584-2018 applicability window, the engine uses the calculated incident-energy method. The input set includes nominal voltage, available bolted fault current, electrode configuration, conductor gap, working distance, enclosure dimensions when known, and protective-device clearing time. The calculation evaluates the full arcing-current branch and the reduced arcing-current branch because the lower current can produce a slower trip and the higher incident energy. The result records which branch governed.

The calculator does not translate calculated incident energy into an NFPA 70E PPE category. Under the incident-energy method, it reports incident energy at the stated working distance and the matching minimum arc-rating basis. A qualified person selects the complete arc-rated clothing system, shock protection, face and head protection, hearing protection, and task controls.

When an input is outside the calculated-method window, the engine considers the NFPA 70E PPE-category method only where an implemented table row applies and every represented condition can be checked. A table-method result states the category and its table basis. When neither a calculated result nor a verified table row applies, the engine returns an engineering-study-required state and withholds incident energy, boundary, and category. It does not extrapolate through a gap in scope.

Input handling and conservative guards

The server is authoritative for validation. Browser constraints help the user, but a request is accepted only after the same fields pass server-side rules. Numeric inputs must be finite and within the product's stated range. Working distances below the model floor are raised to 305 mm with a visible flag because using a shorter, unsupported distance would imply precision the model does not provide.

The reviewed calculator input scope accepts protective-device clearing times from 10 ms through 2,000 ms. A duration above 2,000 ms is rejected. The application does not silently shorten a user-entered duration and does not describe the 2-second guard as a universal physical limit. Cases outside this product scope require qualified engineering judgment about the equipment, task, protective-device behavior, possible worker movement, and appropriate analysis method.

The guided-input assistant distinguishes values supplied by the user, estimates derived from transformer nameplate data, suggestions from a curated protective-device library, and an explicitly accepted conservative default. Provenance travels into the calculation record. Library suggestions carry their citation and a direction to verify them against the manufacturer's time-current curve. Estimates are not silently substituted for utility or study data.

Result and label semantics

Every new label uses WARNING as the default signal word. IEEE 1584 does not create an automatic DANGER switch at 40 cal/cm², so the engine does not infer one from incident energy. An equipment owner may adopt a different documented hazard-communication policy after its own assessment.

Calculated-method outputs show the incident energy, arc-flash boundary, working distance, and minimum arc-rating basis. Table-method outputs show the PPE category only when the table method assigned it. Unsupported cases show that a study is required instead of printing zeros or plausible-looking placeholders. A historical record created by an older behavior remains versioned; if its semantics are no longer accepted, the interface marks it as legacy and calls for a rerun rather than rewriting the original evidence.

Validation evidence

The automated suite covers domain equations, applicability boundaries, HTTP validation, persistence, authorization, signed downloads, billing gates, document generation, content rendering, structured data, sitemap behavior, and security headers. High-value calculation anchors include the published IEEE 1584 Annex D.1 medium-voltage example and fixed low-voltage regression fixtures. Boundary tests exercise voltage, current, gap, working distance, clearing-time floors and ceilings, both arcing-current branches, table fallbacks, and study-required outcomes.

Generated PDF labels, energized-work permits, and calculation records use golden-master comparisons. A wording or layout change causes those comparisons to fail until the new artifact is deliberately reviewed and accepted. Static analysis, formatting, TypeScript checking, linting, production builds, and server-side tests are release gates. The public corrections and change log explains how safety-significant defects are handled.

Automated tests do not constitute independent professional review. The application exposes a separate qualified-review launch flag. That flag stays pending until a qualified electrical-safety reviewer has examined the implemented equations, applicability decisions, regression anchors, PPE wording, label semantics, and representative PDFs. Commercial delivery is configured to fail closed when required mail or billing credentials are absent.

Review a result

A reviewer should confirm the one-line context and equipment identity first, then trace every input to its source. Verify available fault current at the equipment location, not merely at the transformer. Read clearing time at both calculated arcing currents from the correct device curve or trip unit. Confirm electrode configuration, gap, enclosure, and working distance against the actual equipment and task. Check that the stated calculation method applies and that no warning or provenance flag has been ignored.

Finally, treat the output as one input to the work decision. Equipment condition, maintenance, likelihood of an arc event, shock boundaries, job planning, hierarchy-of-risk-control options, and the justification for energized work sit outside a numeric incident-energy calculation. If the field condition differs from the record, stop and reassess rather than editing the label in isolation.