When an overseas project encounters a suspected heat transfer fluid batch issue, the key question is not how quickly the supplier can send an “explanation letter.” The real test is whether the event is identifiable, contained, traceable, verifiable and formally closed.
A defensible supplier investigation should answer six questions: What is the nonconformity? Which lots, quantities and customers may be affected? Are the samples and test data reliable? Is the root-cause evidence sufficient? Have corrective actions been implemented? Has their effectiveness been verified?
Scope note: This article provides a general B2B supplier-audit framework. It does not replace contract specifications, applicable laws, project procedures or agreed third-party inspection requirements.
Descriptions such as “darker color,” “slow heating,” “unusual pump noise” or “different from the previous batch” are investigation clues, not proof of a manufacturing defect. The record should identify the product, grade, lot number, manufacturing date, package and quantity; where the issue was detected; the relevant parameter; the governing specification; and the actual operational or commercial impact.
Color and odor alone should not determine fluid rejection or supplier liability. Temperature history, oxygen exposure, water, contamination, mixing and sampling practice can all affect fluid condition.
Place the suspect lot on hold and stop further shipment, repacking or charging.
Reconcile stock, goods in transit, unused site inventory and quantities already charged.
Use clear status identification and restrict unauthorized release.
Preserve original packaging, seals, labels, receiving documents and representative samples.
Define who may approve retesting, replacement, continued use or disposal.
Containment is not an admission of liability. Its scope should reflect severity, traceability, contractual requirements and documented risk assessment.
Backward traceability should cover critical raw-material lots, suppliers, incoming inspection, production vessel or equipment, charging and process records, filtration, filling, packaging materials and label versions.
Forward traceability should cover finished quantity, retention samples, inventory, delivery documents, transport details, customer and project location, arrival date, and quantities used or unused.
Lateral traceability evaluates other lots associated with the same raw material, equipment, filling line, time window or test condition. These lots should not automatically be rejected, but the supplier should document whether they were included in the risk review and why they were excluded from the affected-lot boundary.
When the supplier’s retention sample passes but the customer sample fails, the audit should compare the sample chain and method chain before selecting either report. Confirm the lot identity, package integrity, sampling point and time, sample homogenization, container cleanliness, transport exposure, test method, instrument, units, temperature conditions and acceptance criteria.
If a third-party retest is required, agree in writing on sample splitting, sealing, numbering, witnesses, laboratory, test method, specification and rules for using the result.
A certificate of analysis can support batch release, but it does not replace manufacturing records and investigation evidence. Verify that the COA matches the contract, lot, quantity and issue date; that values are actual results rather than only “PASS”; that results trace to sample IDs and instrument records; and that out-of-specification results, unusual trends or repeated tests are properly explained.
A passing retention-sample result proves only that the tested sample passed under the stated method. It does not automatically invalidate the customer’s result.
Statements such as “operator error,” “transport problem” or “customer misuse” are not sufficient root causes. Ask which record proves where the issue occurred, why existing controls failed to detect or prevent it, whether other lots face the same risk, and which hypotheses remain unverified.
Tools such as 5 Why, fishbone diagrams and fault trees are useful only when conclusions connect to batch records, equipment logs, raw-material data, original laboratory data, retention-sample results, photographs, interviews or simulation evidence. If evidence is incomplete, state “most probable cause” or “root cause not fully confirmed” and define the remaining investigation.
Correction addresses the detected nonconformity, such as segregation, retesting, rework, replacement or approved disposal. Corrective action addresses the confirmed cause to reduce recurrence risk. Effectiveness verification uses subsequent lots, trend data, audits or checks to confirm that the action achieved its objective.
“Personnel training completed” proves an activity was performed; it does not by itself prove effectiveness. The verification method, scope, timing and acceptance criteria should be defined before closure.
Use one event number and keep product names, grades, lot numbers, quantities, issue descriptions, affected scope, COA/TDS/SDS versions, customer notifications, units, dates, approvers and translation reviews consistent across languages. Commercial compensation should be recorded separately from technical conclusions.
Event number, timeline and acceptance criteria.
Affected-lot, inventory, in-transit and customer-scope lists.
Hold, shipment-stop and status-identification evidence.
Contract specification, COA, sample IDs and laboratory reports.
Agreed retest plan and lot-boundary rationale.
Root-cause analysis with an evidence index.
Corrective-action owners, deadlines and completion evidence.
Effectiveness criteria, results and escalation path.
Bilingual customer communication and final closure approval.
Auditing a Chinese heat transfer fluid manufacturer after a suspected batch issue should move beyond “Did the supplier reply?” to “Does the evidence close the loop?” A reliable sequence is: define the nonconformity, contain the scope, establish three-direction traceability, align samples and methods, explain data differences, determine cause from evidence, implement corrective action and verify effectiveness.
Technical and compliance boundary: This article does not constitute a product conformity decision, legal opinion or assignment of liability. Actual disposition must follow the contract, approved specification, applicable law, project procedure and mutually accepted test results.
