结论：失效分析的价值在于找到原因，而原因决定了后续怎么处理

产品在使用中发生失效——车架断裂、脚踏脱落、制动失效、电池异常——处理的第一步不是赔偿或更换，而是搞清楚为什么。

因为原因不同，后续的处理完全不同：

材料或工艺缺陷。 可能是批次性问题，需要排查同批产品。

设计问题。 所有产品都有风险，可能需要召回或改进。

使用不当或超载。 属于使用问题，但也要看资料中的警示是否充分。

正常磨损到寿命。 属于维护问题，要看更换周期的建议是否合理。

外力或意外。 与产品质量无关。

搞不清原因就处理，要么承担了不该承担的责任，要么放过了真实的风险。

分析的步骤

第一步，保全现场与样品。 不要在分析前清洗、修理、拆解。

第二步，收集信息。 使用者、使用历史、失效时的情况、环境条件。

第三步，外观检查与记录。 拍照、记录失效部位和形态。

第四步，提出假设。 可能的失效机理有哪些。

第五步，选择分析方法。 针对假设选择能验证或排除的方法。

第六步，实施分析。 由浅入深，先做非破坏的。

第七步，形成结论。 根本原因是什么。

第八步，确定措施。 针对原因的纠正和预防。

第一步最关键也最常被破坏。 现场人员出于好意把断裂的部件清洗干净、或者把散落的零件重新装上，这些动作可能破坏了关键证据。

样品与信息的保全

样品保全的要求：

不清洗、不打磨、不修理；断口不要接触和碰撞；妥善包装避免运输中损伤；如果有腐蚀，要防止继续腐蚀；相关的配件、紧固件一并保留；以及记录样品的取得过程。

「断口不要接触」很重要。 断口的微观形貌能说明失效机理，用手摸、用工具刮、或者把断面对合，都可能破坏这些信息。

信息收集的内容：

使用者的身体条件与使用方式；产品的使用时长与强度；失效发生时的具体情况；有无异常的前兆；维护与维修的历史；以及是否有超出说明书的使用。

「有无前兆」的信息很有价值。 失效前是否有异响、松动、变形，这些信息能指向失效的发展过程。

常用的分析手段

手段 能回答什么
宏观检查 失效位置、形态、变形情况
断口分析 失效机理（疲劳、过载、脆断）
材料成分分析 材料是否符合规定
金相或显微结构 材料状态、热处理、缺陷
硬度与力学测试 材料性能是否达标
尺寸测量 是否符合图纸
涂层与表面分析 表面处理的状态
腐蚀产物分析 腐蚀类型与介质

「断口分析」是失效分析中信息量较大的手段。 疲劳断口、过载断口、脆性断口的形貌特征不同，能直接指向失效机理，而机理指向原因。

分析顺序建议由非破坏到破坏。 先做宏观检查和拍照，再做可能改变样品状态的分析。

常见的失效机理

疲劳。 反复加载导致的裂纹萌生与扩展。断口有特征形貌。这是辅具中最常见的机理。

过载。 一次性的超出承载能力。断口呈塑性变形或脆性特征。

腐蚀。 环境作用导致的材料损失或性能下降。

腐蚀疲劳。 两者叠加，寿命明显低于单独作用。

磨损。 摩擦导致的材料损失。

老化。 材料随时间的性能下降，塑料件常见。

制造缺陷。 焊接缺陷、夹杂、气孔、加工痕迹造成的应力集中。

「疲劳起源于应力集中处」是一个有用的判断线索。 如果断口的疲劳源在一个孔边、一个尖角、或者一个加工划痕处，指向的是设计或制造问题。

结论的用途

责任认定。 判断是产品问题还是使用问题。

同批排查。 如果是批次性问题，排查范围。

设计改进。 针对根本原因的改进。

工艺改进。 制造环节的改进。

资料修订。 如果是警示不足，修订说明书。

风险文件更新。 新发现的失效模式纳入风险分析。

上报义务。 涉及严重伤害的事件可能有报告义务。

「风险文件更新」容易被忽略。 实际发生的失效模式如果在原风险分析中没有识别，说明风险分析不完整，应当补充。

分析的局限

要承认分析的局限：

样品被破坏后信息有限。 清洗、修理会丢失证据。

使用历史不清楚时难以定论。 缺少关键信息。

多因素叠加时难以分清主次。 比如既有材料缺陷又有超载。

某些机理需要特殊设备。 不是所有分析都能做。

结论应当与证据相称。 证据不足时给出可能的原因和不确定性，而不是勉强给一个确定结论。

分析报告的内容

失效分析报告建议包含：

委托信息与样品信息。 样品的来源、状态、接收时的情况。

背景信息。 失效的经过、使用历史。

观察记录。 外观检查的结果、照片。

分析过程。 用了什么方法、得到什么数据。

分析与讨论。 从数据到结论的推理过程。

结论。 失效机理与可能的原因。

不确定性说明。 证据的局限、无法确定的部分。

建议措施。 针对原因的建议（如果委托方需要）。

「不确定性说明」体现报告的严谨性。 把能确定的和不能确定的分开说，比给一个看似确定但证据不足的结论要负责。

预防性的应用

失效分析不只用于事后处理，也可以主动使用：

对试验中的失效做分析。 耐久试验中断裂的样品，分析失效位置和机理，指导设计改进。

对退役产品做分析。 看长期使用后哪些部位接近失效。

对竞品做分析。 了解同类产品的薄弱环节。

这种主动应用的价值在于：在失效造成后果之前发现问题。

我们的做法

承接失效分析委托时，我们会先了解失效的经过和样品的处置情况。如果样品已经被清洗或修理，我们会说明这对分析的影响 ——有些信息可能已经无法获得。

分析顺序上，我们从非破坏的宏观检查和记录开始，逐步深入到可能改变样品状态的分析，并在每一步保留记录，便于后续复核。

结论方面，我们会把观察到的事实与推断分开表述，证据支持到什么程度就说到什么程度。

如果你有产品失效需要做原因分析，建议先保全样品不做处理，然后联系：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. When a product fails in service, whether a frame fractures, a footplate detaches, a brake fails or a battery misbehaves, the first step is not compensation or replacement but establishing why. Because the cause determines everything that follows. A material or process defect may be a batch problem requiring investigation of the whole batch. A design problem puts every unit at risk and may require recall or modification. Misuse or overload is a usage problem, though whether the warnings in the documentation were adequate must still be examined. Normal wear to end of life is a maintenance matter, though whether the recommended replacement interval was sensible must be examined. And external force or accident is unrelated to product quality. Acting without establishing the cause means either accepting liability that is not yours or letting a genuine risk pass.

Steps in analysis. Preserve the scene and the sample, without cleaning, repairing or dismantling before analysis. Collect information about the user, the use history, the circumstances of failure and the environment. Inspect and record visually, photographing and documenting the location and form of failure. Form hypotheses about possible failure mechanisms. Select analytical methods capable of confirming or excluding those hypotheses. Carry out the analysis from non-destructive to destructive. Reach a conclusion on root cause. And determine corrective and preventive measures. The first step is both the most important and the most often compromised: site staff mean well when they clean a fractured component or reassemble scattered parts, and those actions may destroy the key evidence.

Preserving samples and information. For samples: do not clean, abrade or repair; do not touch or knock the fracture surface; package carefully to avoid transit damage; prevent further corrosion where corrosion is present; retain associated accessories and fasteners; and record how the sample was obtained. Not touching the fracture surface matters: its microscopic morphology reveals the failure mechanism, and handling it, scraping it with a tool or fitting the halves back together can destroy that information. For information: the user's condition and manner of use; the product's service duration and intensity; the specific circumstances of failure; whether there were warning signs; maintenance and repair history; and whether use exceeded what the instructions permit. Warning signs are valuable information: noises, looseness or distortion before failure indicate how the failure developed.

Analytical methods. Macroscopic examination establishes the location, form and deformation. Fracture surface analysis establishes the mechanism, whether fatigue, overload or brittle fracture. Composition analysis establishes whether the material conforms. Metallography or microstructural examination establishes material condition, heat treatment and defects. Hardness and mechanical testing establish whether properties meet specification. Dimensional measurement establishes conformity with drawings. Coating and surface analysis establishes the condition of surface treatment. And corrosion product analysis establishes the corrosion type and medium. Fracture surface analysis carries the most information: fatigue, overload and brittle fractures have distinct morphologies pointing directly to the mechanism, and the mechanism points to the cause. Proceed from non-destructive to destructive, beginning with macroscopic examination and photography before anything that may alter the sample.

Common failure mechanisms. Fatigue, from crack initiation and propagation under repeated loading, with characteristic fracture morphology; this is the commonest mechanism in assistive products. Overload, from a single excursion beyond capacity, with ductile or brittle fracture characteristics. Corrosion, from environmental attack causing material loss or property reduction. Corrosion fatigue, where the two combine and life falls well below either acting alone. Wear, from friction removing material. Ageing, from property decline over time, common in plastics. And manufacturing defects, where welding defects, inclusions, porosity or machining marks concentrate stress. Fatigue originating at a stress concentration is a useful clue: where the fatigue origin lies at a hole edge, a sharp corner or a machining scratch, it points to a design or manufacturing problem.

Uses of the conclusion. Attributing responsibility between product and usage. Investigating the batch where the problem is batch-related. Improving the design against the root cause. Improving the process at the manufacturing stage. Revising documentation where warnings were inadequate. Updating the risk file to include newly identified failure modes. And meeting reporting obligations where serious injury is involved. Updating the risk file is easily overlooked: a failure mode that occurs in practice but was not identified in the original risk analysis shows that the analysis was incomplete and should be supplemented.

Limitations. Acknowledge them. Information is limited once a sample has been disturbed, since cleaning and repair lose evidence. Conclusions are hard to reach where use history is unclear. Multiple contributing factors are hard to rank, as when a material defect and overload both apply. And some mechanisms require specialised equipment not universally available. Conclusions should match the evidence: where evidence is insufficient, state possible causes and the uncertainty rather than forcing a definite answer.

How we handle it. For failure analysis we establish first what happened and how the sample was handled. Where a sample has already been cleaned or repaired we explain the effect on the analysis, since some information may no longer be obtainable. We work from non-destructive macroscopic examination and recording through to analyses that may alter the sample, keeping records at each step for later review. In the conclusion we separate observed fact from inference and state no more than the evidence supports.

If you have a product failure needing analysis, preserve the sample without treating it and then get in touch: +86 132 4819 8029.