结论：出厂检验的项目要挑「会变的、影响安全的、事后难补救的」
出厂检验（也叫交收检验）是产品出厂前的最后一道关。
它的定位与型式检验不同：
型式检验 ——验证设计，项目全面，周期长，样品少，通常是破坏性的。
出厂检验 ——验证本批（本台）的制造质量，项目精简，必须快速且非破坏。
所以出厂检验不可能覆盖型式检验的全部项目，它要挑。
挑的原则是三条：
会变的。 逐台可能不同的特性——装配质量、功能动作、电气参数。
影响安全的。 制动、电气安全、结构连接。
事后难补救的。 出厂后发现成本高的。
反过来，由设计和材料决定、批内一致的特性（比如材料成分、疲劳寿命、阻燃性能）不适合逐台检，应当在型式检验和进料检验中控制。
必检与抽检的划分
| 类别 | 项目举例 | 方式 |
|---|---|---|
| 逐台必检 | 外观、标识、功能、电气安全 | 每台 |
| 逐台必检 | 制动功能、紧固状态 | 每台 |
| 抽样检验 | 静态稳定性、主要尺寸 | 按批抽 |
| 抽样检验 | 制动距离、噪声 | 按批抽 |
| 不做 | 疲劳、阻燃、材料成分 | 型式与进料控制 |
「逐台必检」的项目要控制数量 ——每台都做，项目多了成本高、节拍慢，应当集中在安全和功能上。
逐台必检的典型项目
外观。 涂层、焊缝、划伤、变形、清洁。
标识。 铭牌、警告标识、序列号的完整与正确。
紧固。 关键螺栓的紧固状态（可用标记法目视确认）。
折叠与展开。 动作顺畅、锁定可靠。
调节功能。 各调节机构的动作和锁定。
制动。 制动的施加与释放、驻车的可靠性。
转动。 车轮转动灵活、无异响。
电气（电动产品）。 开机、调速、电量显示、充电、保护功能。
电气安全（电动产品）。 绝缘或漏电流的快速检查。
随机文件。 说明书、合格证、工具、附件。
「关键螺栓的标记法」是低成本有效的做法 ——紧固后在螺栓和基体上画一条连续的标记线，线错位就说明松动，目视即可确认。
抽样检验的项目
静态稳定性。 按批抽样做倾斜试验。
主要尺寸。 座宽、座深、总宽、总长、折叠尺寸。
质量。 整机重量。
制动距离。 电动产品的制动距离。
速度。 电动产品的最大速度。
续航抽查。 定期抽样验证。
噪声。 运行噪声的抽查。
抽样比例按批量和历史质量水平定，质量稳定可以放宽，出现问题要加严。
判定准则的设定
引用标准。 有标准要求的直接引用。
企业内控。 内控通常严于标准，留出余量。
外观的准则。 外观判定主观性强，建议用实物样板或图片限度样本。
功能的准则。 动作是否顺畅、锁定是否可靠，尽量量化（比如操作力的范围）。
不合格的分级。 严重、一般、轻微，对应不同的处理。
「外观用限度样本」能明显减少争议 ——文字描述「无明显划伤」每个人理解不同，实物样板是统一的尺度。
记录的要求
逐台记录还是批记录。 逐台必检的项目宜逐台记录（至少记录序列号和判定）。
记录的内容。 产品标识、检验项目、结果、判定、检验员、日期。
不合格的记录。 不合格的现象、处理方式、返修后的复检。
放行的凭证。 检验合格后的放行标识。
保存期限。 按规定保存，便于追溯。
「返修后的复检记录」常被遗漏 ——返修后应当重新检验并记录，而不是直接放行。
与型式检验的关系
型式检验验证设计和工艺的能力。 出厂检验验证本次制造的执行。
型式检验的条件。 新产品、设计变更、工艺变更、定期、停产后恢复。
出厂检验不能替代型式检验。 即使出厂检验全部合格，如果设计或工艺变更了，仍需型式检验。
「工艺变更后的型式检验」容易被忽略 ——企业换了供应商或改了工艺，觉得产品没变，但材料和工艺的变化可能影响型式检验项目的结果。
常见的准备不足
信息不全。 材料、工艺、使用环境的信息缺失，导致方案设计缺乏依据。
样品不具代表性。 试制样品或挑选样品，与量产状态有差距。
条件未明确。 测试条件没说清楚，数据的适用范围就不明确。
只看初始值。 忽略了环境和时间作用后的性能变化。
项目不完整。 遗漏了必要的项目，提交时才发现。
这五项在实际委托中出现频率较高，建议在项目启动时逐项确认，因为它们造成的返工成本远高于事前确认所花的时间。
数据的组织与保存
记录的完整性。 条件、方法、设备、样品状态、实测值。
可追溯性。 数据与样品批次、生产批次的关联。
保存期限。 覆盖产品的生命周期。
检索的便利。 需要时能快速定位。
版本的管理。 数据更新时的版本控制。
建议建立统一的数据台账，记录每组数据的来源、条件、适用范围和有效期。数据积累起来之后，后续的产品开发和变更评估都能受益。
与整体工作的衔接
这部分工作不是孤立的，它与几项内容相关：
与设计的衔接。 数据是设计决策的依据，也是设计验证的证据。
与风险管理的衔接。 识别出的风险点应当有对应的验证。
与变更管理的衔接。 变更时判断是否需要重新验证。
与生产控制的衔接。 关键参数的控制限值来自验证数据。
与售后的衔接。 失效分析需要原始数据作为对照。
建议在项目规划时就把这些衔接关系理清楚，避免各环节各做各的、数据无法互相支撑。
我们的做法
企业咨询出厂检验设置时，我们会从检测角度提示哪些项目适合逐台、哪些适合抽样、哪些应当放在型式检验。
抽样项目中的静态稳定性和制动，如果企业没有条件自己做，可以委托我们按批做 ——这类项目设备要求不高但需要规范的方法。
型式检验的触发条件我们会提醒 ——特别是供应商变更和工艺变更，这两类变更后的型式检验容易被漏掉。
如果你在设置出厂检验项目，需要检测角度的参考，或者需要委托批次抽检，可以直接联系：132 4819 8029。检测能力见服务介绍，产品分类见手动轮椅检测与电动轮椅检测，案例见案例。
English version
Conclusion. Final inspection, also called acceptance inspection, is the last gate before shipment. Its purpose differs from type testing. Type testing verifies the design, covering items comprehensively over a long period with few samples, generally destructively. Final inspection verifies the manufacturing quality of this batch or unit, covering a trimmed set of items that must be quick and non-destructive. Final inspection therefore cannot cover everything type testing covers and must select. Three criteria govern selection. Select what varies, meaning characteristics that may differ unit to unit: assembly quality, functional operation and electrical parameters. Select what affects safety: braking, electrical safety and structural connections. And select what is hard to remedy afterwards, meaning what costs a great deal to discover after shipment. Conversely, characteristics determined by design and material and consistent within a batch, such as material composition, fatigue life and flammability, do not suit unit-by-unit inspection and belong in type testing and incoming inspection.
Dividing full from sampled inspection. Unit-by-unit inspection covers appearance, marking, function and electrical safety, together with braking function and fastener condition. Sampled inspection covers static stability and principal dimensions, together with braking distance and noise. And some items are not done at all at final inspection, with fatigue, flammability and material composition controlled through type testing and incoming inspection. The unit-by-unit list must be kept short: doing it on every unit means more items cost more and slow the line, so it should concentrate on safety and function.
Typical unit-by-unit items. Appearance covers coating, welds, scratches, distortion and cleanliness. Marking covers the completeness and correctness of the nameplate, warning labels and serial number. Fastening covers the condition of key bolts, which visual marking can confirm. Folding and unfolding cover smooth operation and reliable locking. Adjustment covers the operation and locking of each adjusting mechanism. Braking covers application and release and the reliability of parking. Rotation covers free running of the wheels without abnormal noise. Electrical function on powered products covers power-up, speed control, charge indication, charging and protective functions. Electrical safety on powered products covers a rapid insulation or leakage check. And accompanying items cover instructions, conformity certificates, tools and accessories. Marking key bolts is a cheap and effective practice: draw a continuous line across the bolt and the base after tightening, and a broken line indicates loosening, confirmable by eye.
Sampled items. Static stability through batch-sampled tilt testing. Principal dimensions including seat width, seat depth, overall width, overall length and folded dimensions. Mass of the whole product. Braking distance on powered products. Maximum speed on powered products. Range, sampled periodically for verification. And running noise, sampled. Set the sampling proportion from lot size and quality history, reducing where quality is stable and tightening where problems have occurred.
Setting acceptance criteria. Cite standards directly where they specify requirements. Set internal limits that are generally tighter than the standard, leaving margin. Use physical or photographic limit samples for appearance, which is highly subjective. Quantify functional criteria where possible, such as ranges for operating force, rather than relying on whether operation feels smooth and locking feels reliable. And grade non-conformances as major, moderate or minor with corresponding responses. Limit samples reduce disputes over appearance noticeably: a written description such as "free from obvious scratches" is understood differently by everyone, while a physical sample is a common yardstick.
Record requirements. Decide between unit and batch records, with unit-by-unit items better recorded per unit, at minimum serial number and judgement. Record product identification, items inspected, results, judgement, inspector and date. Record non-conformances covering the symptom, the disposition and the re-inspection after repair. Provide a release token, being the conformity mark applied after inspection. And retain records for the specified period to support traceability. Re-inspection records after repair are commonly omitted: a repaired unit should be re-inspected and recorded rather than released directly.
Relationship with type testing. Type testing verifies the capability of the design and process, while final inspection verifies the execution of this production. Type testing is triggered by new products, design changes, process changes, periodic requirements and resumption after a production halt. And final inspection cannot substitute for type testing: even with every final inspection passing, a design or process change still requires type testing. Type testing after a process change is readily overlooked: manufacturers change supplier or process and consider the product unchanged, while changes in material and process may affect type testing results.
How we handle it. Where manufacturers consult us on final inspection we offer a testing perspective on which items suit unit-by-unit inspection, which suit sampling and which belong in type testing. Static stability and braking among the sampled items can be commissioned from us batch by batch where the manufacturer lacks the means, both being items of modest equipment demand that require disciplined method. And we raise the triggers for type testing, supplier changes and process changes especially, since type testing after those two is readily missed.
If you are setting up final inspection items or need batch sampling commissioned, get in touch: +86 132 4819 8029.