A filter press pilot run can produce a clean set of numbers, but a number from a pilot is not the same thing as a commitment a supplier can be held to during factory and site acceptance. The gap between the two is where many acceptance disputes start, not because the pilot was wrong, but because the basis behind each pilot result was never carried forward into the contract. Before a buyer sends a pilot report to legal or procurement for conversion into acceptance language, it is worth asking which findings are actually fit to become project commitments, and which are only descriptive of a single test run.
Pilot Findings That Can Become Project Commitments
Not every number a pilot generates belongs in an acceptance clause. A pilot typically reports many observations: cake moisture at a given cycle, filtrate clarity at a given sampling point, cycle time under a given conditioning dose, feed pump behavior at a given pressure. Some of these connect directly to what the buyer is purchasing the equipment to achieve; others are incidental to how the pilot was run and do not generalize past that specific test.
The filter press pilot test acceptance criteria that hold up in a contract are the ones tied to an outcome the buyer actually cares about operating against: processed solids volume or throughput, cake handling characteristics that affect downstream transport or disposal, the quality or destination of the filtrate, the cycle behavior that determines how the equipment will be operated day to day, or an agreed mode of equipment operation such as a stated chamber configuration or automatic cycle sequence. A pilot result that cannot be traced to one of these outcomes may still be useful engineering information, but it is not, by itself, a candidate for a contractual acceptance line.
This distinction matters because a pilot report often mixes both kinds of data in the same table. A reading that looks like a specification is not automatically a basis for acceptance unless it maps onto something the buyer will measure, operate against, or depend on commercially. Where a pilot finding maps cleanly onto one of these buyer-relevant outcomes, it becomes a reasonable starting point for drafting an acceptance term. Where it does not, elevating it to contract language risks creating an obligation around a figure that was never meant to represent full-scale, continuous operation.
The practical judgment for the buyer’s project team is to screen the pilot report against the operating decisions the plant will actually make once the press is installed: what gets measured on shift, what gets reported to environmental compliance, what gets fed into downstream handling. Findings that align with those decisions are the ones worth carrying into acceptance negotiation. Findings that do not align may still inform equipment selection and configuration, but asking a supplier to warrant them turns a design input into an unsupported promise.
Feed Bounds and Configuration Behind Each Metric
| Applicability item | Project-specific detail to fix | Acceptance use |
|---|---|---|
| Feed bounds | Documented feed bounds used for the pilot result | Defines the feed range within which the result may be translated |
| Configuration | Documented configuration behind the metric | Limits the criterion to the agreed configuration basis |
| Conditioning | Documented conditioning used to generate the result | Keeps the criterion tied to the conditioning basis |
| Sampling point | Agreed sampling point for the metric | Keeps acceptance evidence aligned with the pilot sampling basis |
| Measurement method | Agreed method used to determine the metric | Prevents a result measured on a different basis from being treated as equivalent |
Every pilot result was generated under a specific feed and configuration, and that context is what makes the result meaningful or meaningless once it is detached from the pilot and placed into a contract. A cake moisture figure obtained at one solids concentration, one particle size distribution, or one chemical makeup does not automatically apply when the feed shifts outside those bounds. The same is true for configuration: a cycle time or throughput figure generated with a particular chamber arrangement, membrane or recessed-plate configuration, or a particular conditioning dose is a statement about that configuration, not about the filter press category in general.
This is the point where a pilot number quietly becomes a liability if the feed bounds and configuration are not written down alongside it. If the project feed later arrives with a different solids loading, a different fines content, or a different chemical conditioning regime than the pilot used, the original pilot result no longer describes the system the buyer is accepting. Whether that shift matters depends on how far outside the documented bounds the actual feed falls, and whether the supplier’s equipment configuration was selected to tolerate that range or was tuned narrowly to the pilot sample.
Conditioning deserves particular attention because it is often the least visible variable in a pilot summary. A dose of flocculant or coagulant applied during a bench or vendor pilot is rarely the same dose that a full-scale dosing system will deliver under continuous operation, and small differences in conditioning chemistry or dose rate can change cake structure and filtrate clarity independently of the press itself. If the acceptance criterion does not specify the conditioning basis, a result achieved under a favorable conditioning regime cannot be fairly compared to full-scale output run under different dosing.
Sampling point and measurement method complete the applicability record. A filtrate quality figure taken immediately at the discharge differs from one taken after further settling or blending, and a method used to determine moisture or solids content in one lab may not match the method used by another. The EPA Biosolids Technology Fact Sheet on recessed-plate filter presses describes representative bench or vendor pilot testing alongside cake, filtrate, cycle, feed-pump, chamber, cloth, and discharge interfaces, within a municipal biosolids scope; it does not establish industrial-slurry acceptance values, which is precisely why project-specific agreement on these interfaces remains necessary for industrial feed. The buyer’s task is to confirm, before translating any pilot figure into an acceptance number, that the feed bounds, configuration, conditioning, sampling point, and measurement method used in the pilot are documented and that the project feed is expected to stay within them.
Measurement Methods, Calculation Bases, and Tolerances
| Acceptance term | What must be agreed | Decision it controls |
|---|---|---|
| Measurement method | Method used to determine the metric | Whether measurements use the same agreed method |
| Calculation basis | Basis used to calculate the result | How recorded data becomes the acceptance result |
| Tolerance | Tolerance applied to the acceptance term | How variation against the term is judged |
| Sampling frequency | Frequency for collecting acceptance evidence | How often the metric is checked during the agreed test |
Even when a metric clears the first screen — it matters to the buyer’s outcome, and its feed and configuration basis are documented — it still needs a complete and internally consistent definition before it can function as an acceptance term. A number without an agreed measurement method is not verifiable; a number without an agreed calculation basis is not reproducible; a number without a tolerance cannot be judged pass or fail in any defensible way.
Measurement method sets how a quantity is physically determined: how a sample is taken, how moisture or solids content is measured, how pressure or flow readings are logged. Calculation basis is a separate question: once raw readings exist, there must be agreement on how they are converted into the reported result, for example whether an average, a worst-case reading, or a specific cycle’s data defines the accepted value. Two parties can use the same measurement method and still disagree on the outcome if they apply different calculation bases to the same raw data.
Tolerance is where many acceptance disputes actually originate, because a pilot result reported as a single figure implies a precision the underlying process may not support. A filter press cycle, especially one handling variable industrial feed, will not reproduce an identical result on every cycle even under constant nominal conditions. Where the acceptance term has no tolerance band, a single off-nominal cycle can technically fail the criterion even though the equipment is performing within its intended range. Where tolerance is too wide, the criterion stops functioning as a meaningful commitment. The tolerance should reflect the actual run-to-run variability the pilot demonstrated, not an arbitrary margin added during contract drafting.
Sampling frequency determines how much evidence supports the final judgment. A single reading taken at one point in one cycle carries different weight than a series of readings taken across multiple cycles under the agreed conditions. IEC 62381:2024, which addresses factory acceptance testing, site acceptance testing, and site integration testing, supports agreement on acceptance-test scope, activities, responsibilities, and project-specific test-plan checklists; it is written in an automation-system context, and dewatering capacity, cake moisture, filtrate quality, and cycle values still require separately agreed tests built on that same discipline of defining scope and method before testing begins. The buyer’s confirmation point is whether each proposed acceptance term specifies method, calculation basis, tolerance, and sampling frequency together, since any one of the four left undefined reopens the others to dispute.
Witness Points, Data Records, and Abnormal-Run Treatment
| Control item | Project-specific entry required | What the entry establishes |
|---|---|---|
| Witness point | Agreed witness point and witness responsibility | Who is responsible for witnessing the stated acceptance activity |
| Data record | Agreed record for acceptance data | Which record carries the evidence used for acceptance |
| Interrupted run | Agreed treatment of an interrupted run | How an interruption affects use of that run’s data |
| Abnormal run | Agreed treatment of an abnormal run | How abnormal conditions affect use of that run’s data |
A well-defined metric still needs a defined process for generating and recording the evidence that will be judged against it. Witness points establish who is present, and with what authority, when acceptance data is collected. Where witnessing responsibility is left informal, a test run can produce data that one party considers conclusive and the other considers merely indicative, with no agreed record of who observed what and under what conditions.
Data records are the second piece: raw instrument readings, calculated results, and the context notes describing feed conditions during the run all need to live in a record that both parties recognize as the acceptance record, rather than in informal logs that can be contested after the fact. Where the acceptance agreement does not specify which record is authoritative, a dispute over results can become a dispute over which data set even applies.
Interrupted and abnormal runs are the area most often left undefined until a problem actually occurs. A run can be interrupted by a utility fault, a feed supply gap, or an operational pause unrelated to the press’s own performance; an abnormal run can occur where feed conditions drift outside the documented bounds mid-test, or where a conditioning dose is not delivered as intended. If the acceptance agreement is silent on how such runs are treated, the parties are left negotiating after the fact over whether the data from that run counts, is discarded, or triggers a repeat test — exactly the kind of dispute that a documented rule, agreed before testing begins, is meant to prevent.
The practical content of this control is not complicated: decide, before the test campaign starts, what counts as an interruption versus an abnormal run, who has the authority to make that call during the test, and what happens to the data from an affected run. Where that decision is made in advance, an abnormal run becomes a procedural event with a known consequence. Where it is not, the same event becomes a point of contention that can delay acceptance regardless of how the equipment actually performed.
Change Control for Feed or Project Revisions
A pilot-derived acceptance criterion is only as reliable as the assumption that the project feed and configuration will match what the pilot tested. Projects change between pilot testing and commissioning: a feed source is substituted, upstream processing shifts the particle size distribution, a conditioning chemical is changed for cost or availability reasons, or the plant’s throughput target is revised. None of these changes make the original pilot work worthless, but each one can move the actual operating basis outside what the acceptance criterion was built to cover.
The practical risk is not that feed or configuration changes happen — projects evolve, and industrial feed streams are rarely perfectly stable — but that a changed basis gets treated as if it were still covered by the original benchmark. A cake moisture or cycle time figure negotiated against one feed composition does not automatically remain valid once that composition shifts materially, and treating it as if it does transfers the risk of the change onto whichever party is less able to detect or document it.
A change control rule addresses this directly: where the feed or project configuration changes materially from what the pilot documented, that change triggers a review of the acceptance criteria rather than leaving the original numbers in force by default or voiding them without discussion. This is a deliberately narrow mechanism. It does not require renegotiating the whole contract every time a minor variation occurs; it requires an agreed definition of what counts as a material change to feed or configuration, and a process for revisiting the specific acceptance terms that depended on the original basis.
Without this rule, two outcomes are both possible and both undesirable. The acceptance criterion can be applied rigidly to a feed it no longer describes, penalizing a supplier for not meeting a benchmark that was never built for the new conditions. Or the criterion can be treated as automatically void the moment any feed detail shifts, leaving the buyer with no enforceable commitment at all. A documented change rule, agreed at the same time as the acceptance criteria themselves, keeps a revised feed or configuration inside a review process instead of inside a dispute.
Acceptance Record Linking Evidence to Contract Scope
| Record section | Evidence or term to link | Contract-scope connection |
|---|---|---|
| Metric and project outcome | Acceptance metric tied to processed solids, cake handling, filtrate destination, cycle behavior, or agreed equipment operation | Identifies the project outcome being judged |
| Applicability basis | Feed bounds, configuration, conditioning, sampling point, and measurement method | States the conditions within which the pilot benchmark applies |
| Calculation terms | Calculation basis and tolerance | Defines how the acceptance result is determined and judged |
| Test execution | Sampling frequency, witness points, witness responsibility, and data record | Identifies how acceptance evidence is collected and recorded |
| Run exceptions | Agreed treatment of interrupted and abnormal runs | Defines how affected run data is treated |
| Change control | Agreed rule for material changes in feed or project configuration | Sends a changed basis to review instead of silently applying the original benchmark |
By the time a project reaches contract execution, the individual pieces — which metrics qualify, what bounds and configuration they depend on, how they are measured and calculated, how runs are witnessed and recorded, how changes are handled — need to come together in a single record that ties each piece back to what the contract actually promises. This record is what procurement and legal teams will reference when a dispute arises, and its completeness at signing determines how defensible the acceptance process is later.
The project information a buyer supplies for this record is also what enters Porvoo’s configuration and quotation review, since the equipment configuration, conditioning interface, and chamber or plate arrangement proposed for a given feed depend on the same bounds and outcomes that the acceptance record documents. A record built on vague or incomplete project inputs produces an acceptance basis that is equally vague, regardless of how carefully the contract language is drafted around it.
Each acceptance metric in the final record should trace to a stated project outcome — processed solids, cake handling, filtrate destination, cycle behavior, or an agreed mode of equipment operation — rather than standing alone as an isolated figure. Each metric should carry its applicability basis forward: the feed bounds, configuration, conditioning, sampling point, and measurement method under which it was derived. The calculation basis and tolerance should be stated explicitly enough that two parties reviewing the same raw data would reach the same conclusion. Execution details — sampling frequency, who witnesses the test, and which record is authoritative — should be fixed before testing begins, not negotiated during it. The treatment of interrupted and abnormal runs should already be agreed, so that an unusual run during acceptance testing is a procedural matter rather than a fresh negotiation. And the change control rule should be present as a standing term, so that a feed or configuration shift found later is handled as a review rather than as a silent continuation of, or break from, the original benchmark.
Where a project’s pilot data was generated under conditions closely matching the intended full-scale feed, this record mainly confirms and formalizes what the pilot already supports. Where the pilot feed diverges from the anticipated project feed in composition, concentration, or conditioning, the record should flag that gap explicitly rather than let a contract term imply equivalence that the pilot data does not establish. The equipment and interface decisions that follow — including how the filter press integrates with upstream feed and conditioning and downstream cake handling — depend on this record being specific enough to support both configuration review and the acceptance testing that follows commissioning.
Frequently Asked Questions
Q: Should every strong pilot result become a contractual acceptance criterion?
A: No. Select results that measure an agreed project outcome, such as processed solids, cake handling, filtrate destination, cycle behavior, or equipment operation, and retain only those supported by the documented pilot conditions.
Q: What information should be settled before pilot data is converted into acceptance terms?
A: Fix the applicable feed bounds, equipment configuration, conditioning basis, sampling point, measurement method, and intended project outcome first. Then define the calculation basis, tolerance, sampling frequency, witness responsibility, and acceptance data record.
Q: Can a project test result be compared directly with the pilot result if the measurement method changes?
A: Treat the results as non-equivalent until the measurement method, sampling point, and calculation basis have been reconciled. The acceptance record should show that both results were determined on the agreed basis before a pass-or-fail judgment is made.
Q: How should an interrupted or abnormal acceptance run be handled?
A: Apply the treatment agreed before contract execution. Record the interruption or abnormal condition, identify which data it affected, and use the predefined rule to decide whether that run’s evidence can support the acceptance decision.
Q: What should happen if the feed or project configuration changes after the pilot?
A: Trigger the agreed change-review rule before applying the original benchmark. Identify which acceptance metrics and applicability conditions are affected, then confirm the revised test basis instead of carrying the pilot result forward unchanged.
