Read a Heat-Treatment Chart Against the Agreed Process

A smooth time-temperature curve can belong to the wrong furnace run, show only the programmed setpoint, or hide an interval when the recorder stopped collecting data. Its shape alone does not prove that the delivered undercarriage parts followed the approved heat-treatment process. Review starts by linking the chart to the batch, then identifying what each trace actually measured.

This article focuses on documentary evidence. The purpose and metallurgy of the process remain with the existing overview of induction hardening for track rollers. The purchase order, material or product specification, and approved heat-treatment procedure control the live decision. No universal temperature, holding time or cooling rule is supplied here, and the chart does not replace hardness, mechanical or metallographic results.

Link the chart to the correct batch

Begin with the supplier and processing site, purchase order, part number, part revision and material designation. Record the material heat, production lot or serial identifiers and the quantity represented. Then connect those parts to a heat-treatment batch or load ID, furnace run, run date and furnace ID. If the supplier uses several identifiers, preserve their relationship rather than selecting one convenient number.

The batch traveller, furnace log and chart should agree. A chart filename that contains the part number or processing date is not sufficient identity because filenames can be copied, renamed or generated manually. The chart itself or its controlled report needs a stable reference, version and link to the load. If the document was reissued, keep the superseded version and identify what changed.

Record the heat-treatment procedure and revision used for the run. Compare that revision with the one required by the purchase order or approved technical record. A curve that resembles the expected cycle cannot correct a procedure-revision mismatch. If the report and traveller identify different revisions, leave the process basis unresolved until the authorized owner provides a disposition.

Identify the operator, report preparer and required reviewer or approval role. A signature is useful only when the record shows what was reviewed and the signer held the applicable authority. Do not infer approval from a typed name or a chart-export timestamp.

Where test coupons or witness samples are required, record their IDs and locations within or relative to the load. Connect them to the same material, batch and run. A coupon result from another load cannot silently fill a gap in the delivered batch record.

Sulzer’s older company heat-treatment report specification is one example that links purchase-order and material requirements, loads, calibrated thermoelements or equipment, and report review. It is not a universal specification or evidence for a KTSU process. The actual order documents decide which fields and approvals are required.

End the identity check with a closed chain: order and part → material heat or lot → heat-treatment batch/load → furnace run → original chart or report → required coupons and product results. If a link relies only on date, curve shape or filename, mark it missing.

Separate setpoints from actual measurements

A setpoint is the programmed target or profile sent to the control system. A control thermocouple provides input used by that system to regulate the furnace. Recording thermocouples, load sensors or other independent channels provide additional measured data according to the process and equipment arrangement. These signals can be related, but they are not interchangeable.

Label every trace before interpreting it. Record the channel ID, sensor type, physical location, unit, scale and sampling interval. Identify whether a line is a setpoint, controller process value, furnace-zone record, load thermocouple or another signal. A legend that says only “temperature 1” does not establish whether the trace represents furnace air, a control zone or the material.

A setpoint trace does not prove that the load or every monitored zone reached the required temperature. A controller value does not automatically represent the coldest point in a load. Conversely, a load sensor cannot be interpreted without knowing where it was placed and whether that location was required by the approved procedure. Keep the conclusion within the actual channel evidence.

Record the working-zone and load-placement references needed by the procedure. ASTM’s current public page for A991/A991M-25 describes a standard for furnace temperature-uniformity surveys, including working-zone documentation and comparison with applicable permissible variation. That scope does not mean one production chart qualifies the furnace or proves that a survey was current. The relevant equipment-status record must be reviewed separately.

Check the status of thermocouples, recorder channels and associated equipment against the required records. Capture calibration or verification references, validity dates and any limitations. A channel that appears on a chart but lacks a traceable sensor identity or valid status cannot be treated as verified actual data.

Distinguish original data from a rendered chart. The exported PDF or image may rescale axes, average points or omit detail. Record the source file, export time, software or recorder reference where required, and retention location for the raw data. Compare the rendered chart with the original values when the decision depends on a boundary, short event or missing interval.

Eurotherm’s electronic heat-treatment record overview discusses data integrity, archiving, review and user management from a recorder vendor’s perspective. It does not define required recorder features for every order or prove acceptance of a product. Use the approved procedure and quality system to decide what original data and audit information must be retained.

If the process requires atmosphere, quench, agitation, transfer or other records, link them as separate channels or documents. Do not infer an atmosphere condition or quench result from the furnace temperature curve.

Read time, interruptions and deviations

Establish the chart’s time basis before measuring any interval. Record cycle start, time zone, clock synchronization and chart speed or sampling interval. If the export displays elapsed time, link its zero point to the furnace run. A chart with an unlabeled horizontal axis cannot support an exact duration.

Apply the approved procedure’s definitions for ramping, heating, soak or hold start, hold completion and cooling. Do not begin the hold simply when the setpoint changes or the control channel first crosses a line unless the procedure says that event controls. The required start may depend on one or more actual channels, load sensors or stabilization criteria.

Record the actual hold interval using the governing start and end events and preserve the underlying timestamps. Do not round a borderline duration into acceptance without the approved rounding rule. If channels reach the criterion at different times, use the method specified by the controlled procedure and show which channel determined the start.

Review overshoot and undershoot against the permissible range stated in the order, product specification or approved process. A visually small excursion may still require evaluation, while a larger-looking change on a compressed axis may be within the authorized range. Record the numerical event, duration, affected channels and applicable criterion instead of judging the curve’s appearance.

Search for power failures, recorder restarts, thermocouple breaks, flatlined signals, communication loss, door openings, alarms, manual mode changes, transfers and quench delays where the procedure requires those events. Compare operator notes, alarm logs and raw data with the rendered chart. A smooth line after a blank interval does not restore the missing evidence.

A 2024 Sulzer furnace-survey specification provides one company-specific documentation example that includes furnace, lot or heat identity, time-temperature data, setpoints, thermocouple locations, calibration and deviations. It does not establish universal survey rules or prove that a production batch complied.

For every data gap or excursion, record the start and end, affected channels, observable cause, operator response and parts or loads potentially affected. Separate factual observations from root-cause conclusions. “Channel 3 recorded no values from 14:22 to 14:31” is evidence; “the load stayed within range” requires another valid basis.

An authorized deviation should identify the requirement, event, affected batch or quantity, supporting evidence, approval authority, scope and any expiry or one-time limitation. The reviewer of the chart should not approve the excursion by describing it as minor. If authorization is missing, the status stays open.

Keep the chart with the remaining acceptance evidence

The chart is one element of the heat-treatment record. Link it to the furnace calibration and temperature-uniformity survey status required by the process, along with thermocouple and recorder status. Record the applicable working zone and whether the production load was placed within it. Do not claim that the production curve itself qualifies furnace uniformity.

Connect the run to the material certificate and to hardness, mechanical or metallographic results required by the product specification. Preserve coupon or sample traceability. A chart that follows the approved cycle does not prove the part achieved the required hardness, case depth, microstructure or strength. Those conclusions need their own valid tests.

Review nonconformance and deviation records for the first and last potentially affected lot. If equipment status expired, a channel failed or the traveller identifies another run, define the affected scope through the authorized quality process. Do not generalize one accepted disposition to later batches.

Batch and process link	Chart channels and time evidence	Equipment and product records	Status and owner
PO, part, heat, lot, load, run, furnace and procedure revision agree	Setpoint, control and required load channels identified; timestamps and original data retained	Current equipment-status references and required product results linked	Normal: evidence aligns for authorized review
Part and date appear, but heat-treatment load ID is absent	Export shows only a setpoint line; no actual channel identity or time basis	Recorder status not linked	Missing: supplier/process owner supplies original identity and actual records
Traveller identifies run R-18; chart header and report identify run R-81	Both files show plausible curves under the same procedure	Delivered-lot link cannot be reconciled	Conflict: quality owner resolves which run applies; neither curve is assumed correct
Load and procedure identified	Required channel has a data gap during the claimed hold; later curve is smooth	Alarm note exists, but no authorized deviation is linked	Unresolved: preserve gap and assign deviation review; no inferred pass

The chart-to-batch record should name the reviewer and disposition owner, but it should not grant lot-release authority. Keep the original data, rendered chart, report version, equipment-status references, material and product-test evidence, and every deviation together for the required retention period.

A defensible review does not ask whether the graph looks like a heat-treatment cycle. It asks whether the correct batch is linked to the correct procedure and furnace run, whether the traces are actual identified measurements, whether the controlled time and deviation rules were applied, and whether the remaining equipment and product evidence supports the separate conclusions required by the order.