APQP PPAP PLM: Digital Quality Planning, Control Plans, PFMEA Links and Part Submission Warrants

Most production part approval packages are assembled the week before they are due, by a quality engineer who copies characteristic numbers from a ballooned PDF into a spreadsheet, then into a PFMEA, then into a control plan, then into a dimensional report. Each copy is a chance to break the chain. Running APQP PPAP PLM as one connected data model, rather than a folder of documents, is how you stop that.

It matters now because the rulebooks moved. AIAG published the third edition of its Advanced Product Quality Planning manual and a standalone first-edition Control Plan manual in 2024, VDA issued a risk-based seventh edition of its Volume 2 approval process in early 2026, and AIAG and VDA released a harmonised SPC manual in July 2026. Every one of them assumes digital, linked records.

You will leave with a gate model, a characteristic traceability chain, a data model sketch, a worked PPAP example and a failure-mode list.

What this covers: APQP phases as PLM gates, the drawing-to-SPC characteristic chain, the PPAP 18 elements and five submission levels, the part submission warrant, VDA 2 and AS9145 equivalents, supplier portals, change management, and what breaks.

Context and Background

Advanced product quality planning (APQP) is the structured method the automotive industry uses to plan a product and its manufacturing process so that the first saleable part is a good part. AIAG’s model runs through five phases: plan and define the program; product design and development; process design and development; product and process validation; and production, feedback, assessment and corrective action. The production part approval process (PPAP) is the evidence package that closes phase four. It proves that a supplier’s production process, run at rate with production tooling, makes parts that meet every engineering requirement.

Both live under IATF 16949:2016, the automotive quality management system standard, which requires a customer-recognised product approval process (clause 8.3.4.4) and points organisations to their customers’ specific requirements. In practice those requirements name the AIAG core tools: APQP, PPAP (fourth edition, 2006), the AIAG-VDA FMEA Handbook (2019), measurement systems analysis (MSA) and statistical process control (SPC). The AIAG PPAP page also notes that an IMDS material data submission is needed for each part number before PPAP approval can be granted (AIAG PPAP).

The 2024 APQP third edition was the first rewrite since 2008. According to AIAG’s launch material summarised by DQS, it moved the control plan into its own manual, added sections on sourcing, change management, program metrics, risk assessment and part traceability, and introduced gated management with suggested checklists per gate, including review of sub-tier supplier APQP (DQS summary). The Control Plan first edition describes the plan as a living document, adds Safe Launch requirements, and explicitly covers using software to develop and manage control plans.

Outside North America, German OEMs use VDA Volume 2, production process and product approval (PPA). Aerospace uses SAE AS9145, issued November 2016, which defines APQP and PPAP requirements that integrate with AS9100, AS9102 first article inspection, AS9103 and AS9110.

PLM vendors have largely treated all of this as a document-control problem. Our survey of the PLM vendor landscape in 2026 shows quality modules bolted onto the side of product structure. My thesis is different. The unit of APQP is not the document; it is the characteristic. If every characteristic is a first-class object with a stable identity, the PFMEA, control plan, inspection plan, gauge study and capability study become views over one graph, and the PPAP package becomes a query rather than a scramble.

The APQP PPAP PLM Reference Architecture

Direct answer: APQP and PPAP in PLM means running the five APQP phases as gated PLM projects, storing every product and process characteristic as a linked object that flows from the drawing or model through DFMEA, PFMEA, control plan, inspection, MSA and SPC, and generating the PPAP package and part submission warrant as a baseline query over that graph.

Figure 1: APQP phases modelled as PLM project gates. Each gate checks deliverables that are linked objects in the product and process structure, and gate four releases a frozen PPAP baseline that produces the part submission warrant.

Read the figure top to bottom. A program is created from a customer RFQ or sourcing award. Gate 0 is the program launch check that the third edition added. Each subsequent phase produces deliverables (design records, DFMEA, process flow, PFMEA, control plan, MSA plan, capability studies) that are not attachments but managed objects with lifecycle states. Gate four, product and process validation, freezes a baseline from which the PPAP package is assembled and the part submission warrant (PSW) signed. Phase five keeps the loop alive with SPC signals, field returns and change requests feeding back into the same objects.

APQP phases are project gates, not folders

Most PLM systems already have a project or program object with phases, deliverables and gate reviews. Teamcenter, Windchill, 3DEXPERIENCE and Aras Innovator all ship some form of it. The mistake is to make a deliverable a placeholder that someone satisfies by attaching a file. A gate check that only verifies “a file named PFMEA exists” proves nothing about content, revision or coverage.

The better pattern binds each deliverable to a typed object and a query. “PFMEA complete” becomes a rule: every process step in the released process flow has at least one PFMEA function, every special characteristic in the design record maps to a PFMEA failure cause or failure mode, and no action-priority-High line lacks an owner and due date. The AIAG-VDA FMEA Handbook replaced the old risk priority number with action priority (AP) ratings of High, Medium and Low, which makes this kind of rule straightforward to compute.

That is a direct reading of the third edition’s gated management. It asks for gates aligned to program milestones with a checklist per gate, and it explicitly extends gate reviews to sub-tier supplier APQP. In PLM terms, a supplier’s APQP project is a child project whose gate status rolls up to yours. If your system cannot show “supplier X is at phase three, two PFMEA rules failing”, your gate review is a meeting, not a control.

Gates also need metrics. The third edition added a program metrics section with an example appendix. Useful computed metrics include the share of characteristics with a control method, the share with an accepted MSA study, open AP-High actions by age, and late deliverables by gate. A PLM report can compute all four live because the data sits in one place.

The characteristic is the unit of traceability

A characteristic is a measurable feature of a product or process: a bore diameter, a weld nugget size, a torque, a paint film thickness, an oven temperature. Product characteristics live on the design record. Process characteristics live on the process. Special characteristics, the ones that affect safety, regulatory compliance, fit, function or subsequent processing, carry customer-specific symbols and drive extra controls. IATF 16949 clause 8.3.3.3 requires a multidisciplinary approach to identify them and to document them in drawings, FMEAs, control plans and operator instructions.

Read that requirement as a data-integrity rule. A special characteristic has to appear, with the same identity, in at least four artefacts. If those artefacts are separate spreadsheets, consistency depends on discipline. If the characteristic is one object referenced by four views, consistency is structural.

Figure 2: The characteristic traceability chain. One characteristic identity flows from the model-based definition through risk analysis, control, measurement and monitoring, and SPC results feed back into PFMEA occurrence ratings.

Figure 2 shows the chain I recommend. The design record, ideally a model-based definition (MBD) with semantic product and manufacturing information (PMI) in STEP AP242 or a native CAD format, defines the characteristic and its tolerance. The DFMEA links it to the function and failure it protects. The PFMEA links it to the process step and the failure causes that could violate it. The control plan names the evaluation technique, sample size, frequency, control method and reaction plan. The inspection plan turns that into measurement routines, which the MSA study validates and the SPC system monitors. Capability results then flow back to the PFMEA as evidence for occurrence and detection ratings.

The control plan PFMEA link is the hinge of the chain. The new Control Plan manual mentions “Reverse PFMEA”, where a team walks the shop floor to check that the failure causes in the PFMEA actually match what happens at the station, and family control plans paired with family or foundation FMEAs. Both only work if the link from a control plan row back to a PFMEA cause is a pointer, not a paraphrase.

Ballooning is identity assignment

Ballooning, numbering each requirement on a drawing, is usually treated as a drafting chore for the dimensional report. In this architecture it is the moment a characteristic gets its identity. If the balloon number is generated by PLM, carried as an attribute on the PMI annotation, and written back into the model, every downstream object can reference it. If balloons are drawn on a PDF by a supplier, you have created a second, unmanaged numbering scheme.

The Quality Information Framework (QIF), standardised as ISO 23952, was designed for exactly this hand-off. It carries characteristic definitions, measurement plans and results with persistent identifiers that can link back to model features. Our comparison of STEP AP242, JT and QIF for MBD and PMI covers the format choices. The practical rule is simple: one authority issues characteristic IDs, and it is the PLM system that owns the design record.

Deliverables must bind to configuration

A PPAP is approval of a specific part number at a specific engineering change level, made by a specific process at a specific site. That makes it a configuration problem as much as a quality one. Every object in the package must be pinned to a revision, and the set must be frozen as a baseline. Our guide to configuration management, baselines and effectivity in PLM explains the mechanics.

Without a baseline, the package drifts. A PFMEA edited after submission no longer matches what the customer approved, and nobody can prove which version was in force. With a baseline, the PSW references a baseline ID, and any later change to a member object raises a flag that the approval may need to be revisited.

Walk-through: The PPAP 18 Elements, Levels and a Data Model

This section turns the architecture into concrete objects. It maps the PPAP 18 elements to PLM sources, explains the five submission levels and the part submission warrant, sketches a data model, and walks through a worked example.

The PPAP 18 elements mapped to PLM objects

The AIAG PPAP fourth edition lists 18 requirements. Most quality teams know them as a checklist. The more useful question is which system of record owns each one, and whether it can be generated or must be attached.

#	PPAP element	Owning object in a PLM-centred model	Generated or attached
1	Design records	Released CAD/MBD and drawing at the PPAP revision; ballooned characteristic list	Generated from design record
2	Authorized engineering change documents	ECNs not yet incorporated into the design record	Query on change objects
3	Customer engineering approval	Customer approval record, when required	Attached evidence
4	Design FMEA	DFMEA object, if the organisation is design-responsible	Generated report
5	Process flow diagram	Manufacturing process plan (MBOM routing)	Generated from process plan
6	Process FMEA	PFMEA object linked to process steps	Generated report
7	Control plan	Control plan object linked to PFMEA and characteristics	Generated report
8	Measurement system analysis studies	MSA study objects per gauge and characteristic	Generated summary plus raw data
9	Dimensional results	Inspection results keyed by characteristic ID	Generated from QIF or CMM results
10	Records of material and performance test results	Lab test results tied to specification clauses	Attached or integrated from LIMS
11	Initial process studies	Capability studies on special characteristics	Generated from SPC data
12	Qualified laboratory documentation	Lab scope and accreditation records	Attached, reused across parts
13	Appearance approval report	AAR, for parts with appearance requirements	Attached with customer sign-off
14	Sample production parts	Sample shipment record and labels	Logistics record
15	Master sample	Retained sample record and location	Logistics record
16	Checking aids	Fixture and gauge records with calibration	Asset records
17	Customer-specific requirements	CSR checklist per customer	Checklist object
18	Part submission warrant	PSW generated from the baseline	Generated, then signed

Two things stand out. First, at least ten of the PPAP 18 elements can be generated from structured data if the characteristic chain exists. Second, the elements that cannot, such as lab accreditation or appearance approval, are mostly reusable evidence with their own lifecycle. A qualified laboratory scope does not change per part; it should be referenced, not re-uploaded.

Submission levels decide what leaves the building

PPAP separates what must be prepared from what must be submitted. A supplier always prepares the full set and retains it. The submission level, chosen by the customer, decides what is sent.

Level	What is submitted to the customer
1	Part submission warrant only (plus an appearance approval report for designated appearance items)
2	PSW with product samples and limited supporting data
3	PSW with product samples and complete supporting data (the default unless the customer specifies otherwise)
4	PSW and other requirements as defined by the customer
5	PSW with product samples and complete supporting data available for review at the supplier’s manufacturing location

In a PLM model the level is a parameter on the submission object, and a rule table maps level to the elements that are packaged versus retained. That rule table should be data, not code, because customer-specific requirements often modify it. For example, some customers require specific elements at level 2 that the manual leaves to their discretion.

Level 5 deserves attention. Nothing is shipped in bulk; the customer audits at the plant. That only works if the full record can be shown quickly and consistently, which is an argument for a live, linked package rather than binders.

The part submission warrant as a computed document

The part submission warrant is the one-page summary that the supplier signs to declare that the parts were made from production tooling and processes and meet all drawing and specification requirements. It carries the part number and engineering change level, the reason for submission, the requested submission level, the part weight, a declaration of results, and the customer disposition.

Most fields on the PSW are derivable. Part number and change level come from the baseline. Reason for submission comes from the triggering event, such as an initial submission, an engineering change, a tooling transfer, a sub-supplier or material source change, or production at an additional location. The declaration of results can be pre-checked by rule: every characteristic has a result, every result is in tolerance, and every special characteristic has a capability study.

The one thing that cannot be computed is the signature. The authorised signatory declares that the evidence is true. A PLM workflow should present the computed checks, block signing if any fail, and record the signature against the baseline hash. The PSW then becomes a cryptographically anchored statement about a known set of objects.

Initial process studies and acceptance thresholds

PPAP requires initial process studies on special characteristics, normally using the performance index Ppk on a significant production run. The fourth-edition acceptance guidance is widely cited as: an index above 1.67 meets acceptance criteria; 1.33 to 1.67 may be acceptable but needs contact with the customer; below 1.33 does not meet criteria. Customer-specific requirements can override these numbers, so they belong in a rule table keyed by customer and characteristic class.

The new harmonised AIAG and VDA SPC manual, on sale from 1 July 2026 according to DQS, reinforces the distinction between capability indices (Cp, Cpk) for statistically stable processes and performance indices (Pp, Ppk) when stability is not demonstrated (DQS on the AIAG and VDA SPC manual). For PLM integration that means a capability study object must record whether stability was assessed, which chart was used, and how many subgroups were collected. Our SPC control charts and Cpk streaming architecture post covers how those studies are computed from shop-floor data.

A data model sketch

The model below is deliberately small. It is not a vendor schema; it is the minimum set of entities and relationships that makes the traceability chain and the PPAP query work. Map it onto whatever your PLM and quality systems already provide.

Figure 3: Minimum data model for digital APQP and PPAP. The characteristic sits at the centre, referenced by risk, control, measurement and capability objects, and a PPAP submission pins all of them to one baseline.

The core entities are these. A part revision owns many characteristics. Each characteristic has a stable ID, nominal, tolerance, unit, classification (standard or special, with the customer symbol) and a pointer to the model feature or PMI annotation that defines it. A process step belongs to a routing on the manufacturing bill of materials and owns process characteristics. FMEA lines link a function, failure mode, effect and cause to characteristics and process steps. A control plan row links one characteristic and one process step to a control method, sample size, frequency and reaction plan, and points back to the PFMEA cause it mitigates.

Measurement objects follow. A gauge or measuring system has an MSA study per characteristic it measures. An inspection result records a measured value against a characteristic ID, a part serial or sample number, and a gauge. A capability study aggregates results for one characteristic over a run and stores Ppk or Cpk with the stability evidence. Finally, a PPAP submission references a baseline, a submission level, a reason for submission, the customer and the generated PSW.

# Illustrative schema sketch, not a vendor data model
Characteristic:
  id: CHR-000184          # issued once by PLM, carried in PMI and balloon
  part_revision: PRT-7731/C
  feature_ref: PMI-annot-0042
  nominal: 12.000
  tol_lower: -0.018
  tol_upper: 0.000
  unit: mm
  classification: special   # customer symbol stored separately
ControlPlanRow:
  characteristic: CHR-000184
  process_step: OP-30
  pfmea_cause: PFM-OP30-C07
  method: in-process air gauge, 100 percent
  reaction_plan: RP-STD-04
PPAPSubmission:
  baseline: BL-PRT-7731-C-PPAP-01
  level: 3
  reason: initial_submission
  psw_status: draft

The rule set that drives gate four is then a handful of graph queries. Every special characteristic must have at least one control plan row, an accepted MSA study for the measuring system named in that row, a complete set of dimensional results on the PPAP samples, and a capability study meeting the customer threshold. Every PFMEA line with AP High must have an action whose status is closed or formally accepted. Every pending engineering change on the part must be either incorporated into the design record or listed as element 2.

Supplier portal collaboration

Most PPAPs flow in one direction: a supplier submits to an OEM or a tier-one. That means the characteristic chain spans two companies and at least two systems. A supplier portal is where the chain either survives or breaks.

Figure 4: Digital PPAP package assembly across the customer and supplier boundary. The customer publishes characteristics with stable IDs, the supplier returns linked evidence, and the PLM validates the package before disposition.

The sequence in Figure 4 has five steps. The customer releases the design record and publishes the characteristic list with IDs, special characteristic classes and the requested submission level. The supplier imports that list into its own quality system, builds its PFMEA and control plan against those IDs, and runs the significant production run. The supplier uploads results, MSA and capability studies keyed by the same IDs, plus attached evidence such as lab reports. The portal runs the same completeness rules the customer would run, so gaps are caught before submission. The customer engineer reviews, sets the disposition (approved, interim approval or rejected) and the PLM records it against the baseline.

The critical design choice is that the customer never accepts characteristic data without IDs. A dimensional report that lists “dimension 14” against a PDF balloon cannot be joined to anything. A QIF results file or a structured CSV keyed by characteristic ID can.

Linking PPAP to engineering change

PPAP is not a one-time event. The fourth edition requires customer notification and, often, resubmission when product or process changes: a design change, a new sub-supplier, a tooling move, a process change, production at a new site. Each of those is an engineering or manufacturing change in PLM.

The link is a trigger rule. When a change object affects a part revision that has an approved PPAP baseline, PLM computes the set of affected characteristics, flags the dependent FMEA lines, control plan rows and capability studies, and proposes a PPAP resubmission with a reason for submission pre-filled. Our post on engineering change management with ECR, ECO, ECN and effectivity covers the change mechanics; the extra field you need is “requires customer notification”, computed from the affected characteristics and the customer’s trigger rules.

This is also where the APQP third edition’s new change management section lands. A change that touches only standard characteristics on a non-safety part may need only notification. A change that touches a special characteristic or the process step that controls it almost always needs a new approval.

Worked example: a machined valve body

Here is an illustrative example. The part, numbers and outcome are hypothetical, chosen to show the mechanics rather than to report a real program.

A tier-two supplier machines an aluminium valve body for a brake module. The released design record at revision C has 112 ballooned characteristics. Eleven are special characteristics, including a 12 mm bore diameter with a tolerance of minus 18 to zero microns (CHR-000184), two sealing surface flatness callouts and a thread depth. The customer requests a level 3 submission.

PLM generates the characteristic list and the supplier portal publishes it. The supplier’s PFMEA covers 9 operations and 64 failure causes. The gate four rule check runs and reports three gaps. First, CHR-000184 is controlled by an air gauge at OP-30, but the MSA study for that gauge was run on a different bore size. Second, one flatness characteristic has no control plan row because the old spreadsheet merged it with a second callout. Third, an ECN for a chamfer change, released after the PPAP samples were cut, is neither incorporated nor listed.

Each gap is a query result, not an auditor’s finding. The supplier reruns the gauge study (illustrative result: the study passes), adds the missing control plan row, and lists the ECN as element 2 with an agreed implementation date. The bore capability study on the significant production run returns an illustrative Ppk of 1.52. That is in the 1.33 to 1.67 band, so the rule flags customer contact rather than blocking. The customer accepts an interim approval with a containment action: 100 percent air gauging continues until a 1.67 index is shown over a longer run.

The PSW is generated from baseline BL-PRT-7731-C-PPAP-01, signed, and dispositioned as interim. The interim status is itself an object with an expiry date, so it cannot quietly become permanent.

Equivalents beyond AIAG: VDA 2 and AS9145

The same architecture serves other approval schemes if the rules are data. The table compares the three most common.

Aspect	AIAG PPAP (4th ed., 2006)	VDA Volume 2 PPA (7th ed.)	SAE AS9145 (2016)
Industry	Automotive, North American OEM-led	Automotive, German OEM-led	Aerospace and defence
Structure	18 elements, levels 1 to 5	Risk-based deliverable list; no submission levels	APQP phases with PPAP as validation evidence
Who decides scope	Customer sets level; CSRs modify	Agreed per deliverable: self-assessment, either, or submission	Customer and contract, with AS9102 FAI integration
Summary record	Part submission warrant	PPA report with self-assessment and customer decision	PPAP approval plus AS9102 first article report
PLM implication	Level-to-element rule table	Per-deliverable disposition flags	Link PPAP to FAI and key characteristics

The VDA 2 seventh edition, dated January 2026 on the VDA QMC download, describes a risk-based approach, fewer required submissions, referencing of existing deliverables, a reduced set of forms and an online PPA tool replacing Excel templates. It also confirms that submission levels and conditional approvals were removed in the sixth edition and that harmonisation with AIAG PPAP is a possible long-term goal (VDA QMC, VDA 2 seventh edition). For PLM this matters because the disposition moves from package level to deliverable level. Your submission object needs a per-element status, not one checkbox.

AS9145 is the aerospace counterpart. SAE describes it as establishing requirements for performing and documenting APQP and PPAP, integrating with AS9100, AS9102, AS9103 and AS9110 (SAE AS9145). Aerospace adds a strong first article inspection culture, so the characteristic ID must also join to AS9102 forms. The good news is that the same characteristic graph serves both.

Trade-offs, Gotchas, and What Goes Wrong

The characteristic-centred model is better than document control, but it fails in specific, predictable ways.

Characteristic ID collisions and re-numbering. When a drawing is revised, drafters often re-balloon from scratch. If the PLM issues new IDs for unchanged features, every downstream link breaks and every capability study looks orphaned. The rule must be that IDs persist across revisions unless the feature itself is deleted, and that a changed tolerance is a new version of the same characteristic, not a new characteristic.

Granularity mismatch. One drawing callout can hide several characteristics, such as a position tolerance with multiple datum references, or a note that applies to “all edges”. Conversely, a shop floor may measure two characteristics with one gauge in one operation. Decide early how many characteristic objects a composite callout produces, and record the rule. Our worked example’s missing control plan row came from exactly this ambiguity.

Supplier system boundaries. Most suppliers will not run your PLM. If the portal only accepts PDFs, you are back to unlinked documents. Require structured results keyed by characteristic ID, and accept that small suppliers will need a light web form or template that enforces the keys. Budget for supplier onboarding; it is usually the longest part of rollout.

Over-automation of judgement. Rules can verify completeness and thresholds. They cannot verify that a PFMEA is thoughtful or that a reaction plan is realistic. A green dashboard can hide a copied foundation FMEA that was never adapted to the actual line. Keep human review at each gate, and use the Reverse PFMEA walk the Control Plan manual describes as a counterweight.

Baseline sprawl. Every interim approval, resubmission and site change creates a baseline. Without clear naming and supersession rules, nobody knows which PPAP is current for a part at a plant. Model a “current approval” pointer per part, site and customer, and make superseded baselines read-only.

Customer-specific requirements drift. CSRs change more often than the manuals. If level-to-element mappings and Ppk thresholds are hard-coded, they will be wrong within a year. Store them as versioned rule tables with an effective date.

Edition transitions. The APQP third edition and Control Plan first edition coexist with control plans written to the second edition; DQS reports that suppliers do not have to update existing control plans to the new requirements. VDA 2 programs started under the sixth edition may finish under it. Tag each project with the edition it follows, so gate rules evaluate against the right version.

Performance and scale. A mid-sized supplier can have tens of thousands of characteristics across active parts and millions of inspection results. The graph links belong in PLM; raw measurement streams belong in a quality or SPC store with characteristic IDs as the join key. Pulling every measurement into PLM is a common way to make the system slow and the project fail.

Practical Recommendations

Start with identity, not workflow. The single most valuable change is to make PLM the issuer of characteristic IDs and to carry those IDs on the model, the drawing balloons and every downstream artefact. Workflow automation built on top of copied numbers just automates errors.

Model the five APQP phases as a project template with gate rules expressed as queries. Begin with three or four rules that catch the gaps you actually see in audits, such as special characteristics without control plan rows, or MSA studies that do not match the gauge named in the control plan. Add rules as you learn.

Treat the PPAP package as a baseline plus a generated report, not a folder. Generate elements that come from structured data, reference reusable evidence such as lab accreditation, and attach only what is genuinely one-off. Make the PSW a computed draft that the signatory approves.

Connect PPAP to change management on day one. Any change that touches a part with an approved PPAP should compute affected characteristics and propose resubmission. This is where most real non-conformances originate.

Finally, design for the supplier. Publish characteristic lists with IDs, accept structured results, and run the same completeness checks in the portal that you run internally.

Checklist:

[ ] PLM issues persistent characteristic IDs and writes them to PMI and balloons.
[ ] Special characteristics are linked to DFMEA, PFMEA, control plan and work instructions.
[ ] APQP project template has Gate 0 to Gate 5 with query-based deliverable rules.
[ ] Submission level and customer thresholds live in versioned rule tables.
[ ] PPAP package is generated from a frozen baseline; PSW references the baseline ID.
[ ] Engineering changes on approved parts trigger impact analysis and resubmission proposals.
[ ] Supplier portal enforces characteristic-ID-keyed results.
[ ] Each project records which APQP, Control Plan, PPAP or VDA 2 edition it follows.

Frequently Asked Questions

What are the PPAP 18 elements?

The PPAP 18 elements in the AIAG fourth edition are: design records; authorized engineering change documents; customer engineering approval; design FMEA; process flow diagram; process FMEA; control plan; measurement system analysis studies; dimensional results; records of material and performance test results; initial process studies; qualified laboratory documentation; appearance approval report; sample production parts; master sample; checking aids; customer-specific requirements; and the part submission warrant. Not every element applies to every part. A supplier prepares and retains all applicable elements, and the submission level determines which are sent.

What is the difference between APQP and PPAP?

Advanced product quality planning is the whole planning method, covering program definition, product design, process design, validation and production feedback. PPAP is the approval evidence produced at the end of the product and process validation phase. APQP asks whether the product and process were planned to be capable. PPAP proves, with parts from a significant production run on production tooling, that they are. In a PLM system, APQP is the gated project and PPAP is the frozen baseline and package that gate four releases.

Which PPAP submission level is the default?

Level 3 is the default under the AIAG PPAP manual unless the customer specifies another level. Level 3 means submitting the part submission warrant with product samples and complete supporting data. Level 1 is the warrant alone, level 2 adds samples and limited data, level 4 is the warrant plus whatever the customer defines, and level 5 keeps the complete data at the supplier’s site for on-site review. Customer-specific requirements frequently adjust these definitions, so check each customer’s manual.

What does a part submission warrant contain?

A part submission warrant summarises a PPAP submission on one page. It identifies the part number and engineering change level, the organisation and manufacturing site, the reason for submission, the requested submission level, the part weight and a declaration that the parts meet all drawing and specification requirements and were made with production tooling and processes. An authorised supplier representative signs it, and the customer records its disposition. In a digital model, most fields come from the baseline and only the signature needs a person.

How does the control plan link to the PFMEA?

Every control plan row should address a characteristic and a process step whose risk is analysed in the PFMEA. The PFMEA identifies failure causes and current controls; the control plan specifies how those controls run in production, including technique, sample size, frequency and reaction plan. The control plan PFMEA link should be a direct reference from the row to the PFMEA cause it mitigates, so that a change in one flags the other. The AIAG Control Plan manual also describes Reverse PFMEA to verify the link on the shop floor.

Can PLM replace a dedicated quality management system for PPAP?

Usually not entirely, and it does not need to. PLM is best placed to own design records, characteristic identity, FMEA, control plans, change and baselines. Measurement data, SPC, calibration and nonconformance handling often live better in a quality management system or manufacturing execution system. The architecture works when characteristic IDs are the shared key and PLM can query capability and MSA results by that key. Replacing everything with one tool tends to create performance problems and push shop-floor users into unfamiliar interfaces.

APQP and PPAP in PLM: Digital Quality Planning, Control Plans, PFMEA Links and Part Submission Warrants (2026)