The short answer
A successful pilot does not automatically define a repeatable production order. Release custom liquid filter bags against a fixed construction revision, documented trial conditions and agreed incoming checks, with an explicit route for deviations and later changes.
Freeze what the pilot actually tested
Give each prototype a unique identity linked to its drawing revision, materials, ring, seam construction and supplier batch reference where available. Record deviations from the proposed specification before installation. An unrecorded 'small improvement' can make the trial impossible to reproduce.
Create a proposed release pack rather than assuming the supplier already operates your preferred quality system. It should contain the drawing, construction schedule, inspection requirements, trial report and purchase-order reference. Pentair's configuration sheet is a useful named example of separating several ordering attributes; its codes do not define a JINLV order.
Separate three acceptance gates
First, check that the supplied articles match the agreed construction and dimensions. Second, confirm safe fit in the identified housing and basket. Third, assess the filtration duty with the process owner. Passing one gate must not silently waive the others. A bag can fit and still fail downstream quality requirements.
Write who records each decision and what constitutes a hold. If a measurement is disputed, review the defined feature, reference points, method and uncertainty before accepting a convenient result. NIST's traceability guidance supports this distinction between a documented measurement result and merely owning a calibrated instrument.
Build a release decision matrix
Use the matrix to make the order status explicit. Sample quantities and sampling plans must be agreed for the risk and process; there is no universal number of pilot bags that proves consistency.
| Gate | Evidence to retain | Release decision |
|---|---|---|
| Prototype identity | Part revision, sample IDs and declared deviations | Hold if the tested construction cannot be identified |
| Fit acceptance | Housing and basket IDs plus observed installation outcome | Hold if seating or support requires improvisation |
| Duty trial | Conditions, downstream result, differential-pressure history and stop reason | Accept only for the documented envelope and agreed criteria |
| Repeat supply | Incoming checks, batch identification and change-notification agreement | Release the defined revision, not an unrestricted equivalent |
Record the operating envelope and unsuccessful results
A trial record should identify liquid condition, temperature, flow, solids description and relevant operating cycles. Record initial and subsequent differential pressure using the agreed system, together with downstream quality observations. Define shutdown and change-out criteria before starting; trained operators must use the approved housing and site procedures.
If the pilot stops early, retain the stop reason and any missing evidence. A run with little contamination may show installation behavior but say little about loading performance. Conversely, a difficult batch should not be erased from the record because another run went well. State what the trial actually covered and what remains untested.
Make the repeat order a controlled decision
The purchase order should reference the accepted drawing and inspection revision, required identification, packaging needs and disposition of nonconforming goods. Agree which records accompany shipments and which remain available on request. Do not claim batch traceability unless the supplier has confirmed the identifiers and record links it can provide.
Define change triggers such as media source, ring material, seam process, dimensions or production location when relevant to qualification. A repeat order after a process-fluid change also needs review, even if the bag is unchanged. Keep accepted and superseded reference samples clearly separated. A second source needs its own evidence; similarity to the approved sample is a starting point, not automatic equivalence.
Your review checklist
- Identify every pilot article and its construction revision.
- Record deviations before the trial begins.
- Assign separate dimensional, fit and duty acceptance gates.
- Agree operating envelope, measurements and stop criteria.
- Define incoming inspection, batch identification and record supply.
- Set change-notification and nonconformance disposition requirements.
Limits of this guidance
A proposed procurement release framework, not a statement that JINLV has performed trials, established lot traceability or obtained engineering approval. Acceptance limits and qualification coverage remain customer-, supplier- and application-specific.
Questions buyers ask
How many successful pilot runs are enough?
There is no universal count. Agree coverage based on process variation, consequences of failure and the evidence needed; do not replace a representative trial plan with an arbitrary number.
Can purchasing accept an equivalent material to avoid a delay?
Only through the agreed change process. Identify the difference, assess its effect on the accepted evidence and obtain the required technical decision before treating it as the released construction.
Technical references
References support general principles or the named supplier's own products; they do not certify or validate a JINLV configuration.
- Liquid Filter Bags Spec Sheet
- Metrological Traceability: Frequently Asked Questions and NIST Policy
- Installation Operation and Maintenance Manual - Eaton Bag Filter Housing (US)
AI-assisted technical synthesis of the references below, with editorial checks. This is general procurement guidance, not an independently signed engineering review or a product validation report.
English is the indexing version. Chinese and Indonesian are reading translations of the same article.