3D 打印器械的检测难点

结论：难点不在能不能做出来，在一致性和内部质量怎么证明

增材制造用于医疗器械已经不新鲜——定制化植入物、手术导板、模型、以及某些结构件。

但它的检测有几个特有的难点：

一致性难以保证。 逐层堆积的工艺，参数的微小波动就可能影响结果，而每件产品可能是不同的（定制化）。

内部质量难以检查。 内部可能有未熔合、孔隙、未去除的支撑或粉末，而这些从外部看不到。

性能的各向异性。 打印方向不同，力学性能不同。

后处理的影响。 热处理、表面处理、支撑去除，每一步都影响最终性能。

一致性的问题

批次的概念模糊。 传统制造的「批次」概念在定制化生产中不直接适用。

参数的敏感性。 激光功率、扫描速度、层厚、粉末状态的微小变化都影响结果。

设备状态的影响。 设备的校准和维护状态。

原料的批次。 粉末的批次差异、回收粉末的使用。

位置的影响。 同一次打印中，不同位置的件性能可能不同。

「同一次打印中不同位置的差异」是这类工艺特有的，成型舱内的温度场和气流分布不均，边缘和中心的件性能可能有差别。

一致性的验证方法

随炉试样。 每次打印同时打印标准试样，用于验证本次打印的材料性能。

位置的覆盖。 随炉试样布置在不同位置。

过程参数的记录。 每次打印的完整参数记录。

过程监测。 打印过程中的实时监测（熔池、层面图像）。

设备的定期验证。 用标准件验证设备状态。

粉末的管理。 批次、回收次数、性能监测。

「随炉试样」是这类工艺一致性验证的核心手段 ——因为产品本身可能不能破坏性测试（定制植入物），随炉试样提供了本次打印的材料性能数据。

内部缺陷的检测

孔隙。 未完全熔合形成的孔洞，影响强度和疲劳。

未熔合。 层间或道间的结合不良。

残留粉末。 内部通道和多孔结构中的未去除粉末。

残留支撑。 内部支撑未完全去除。

检测方法。 工业 CT 是主要手段，能看到内部结构；密度测量可以间接反映孔隙率；金相切片是破坏性的但分辨率高。

「工业 CT」对这类产品价值明显，因为它是非破坏的，能看到内部的孔隙、未熔合、残留粉末，而这些从外部完全看不到。

难点与应对的对照

各向异性

成因。 逐层堆积使层间的结合与层内不同。

表现。 垂直于层面方向的强度通常低于平行方向。

影响。 打印方向的选择影响最终性能。

测试的要求。 不同方向分别取样测试。

设计的应对。 受力方向与打印方向的配合。

「测试应当覆盖不同方向」，只测一个方向的数据不能代表材料的性能，而设计时要知道各方向的性能才能合理布置打印方向。

后处理的影响

支撑去除。 去除后表面的状态、是否有残留、是否损伤本体。

热处理。 消除内应力、调整组织，对力学性能影响很大。

表面处理。 机加工、喷砂、抛光、化学处理。

清洗。 去除粉末和加工残留。

灭菌。 多孔结构的灭菌难度更大。

「多孔结构的清洗和灭菌」是难点 ——粉末可能残留在孔隙深处，清洗液和灭菌介质也可能难以到达，这对多孔骨长入结构尤其需要验证。

表面特性

粗糙度。 增材制造的表面通常较粗糙。

未熔颗粒。 表面附着的部分熔化颗粒可能脱落。

后处理后的表面。 机加工或抛光后的状态。

多孔结构的表面。 孔的尺寸、连通性、表面状态。

清洁度。 表面和孔隙内的残留。

「表面附着的未熔颗粒」是潜在的颗粒来源，它们结合不牢，使用中可能脱落，对植入产品是需要控制的。

验证的组织

工艺验证。 设备、参数、材料的组合验证。

随炉验证。 每次打印的材料性能验证。

首件验证。 新设计的首件全面检验。

抽检。 批量产品的抽样检验。

过程监测。 打印过程的实时数据。

建议把验证重点放在工艺和过程上，因为定制化产品的逐件全检不现实，工艺稳定才能保证产品稳定。

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

承接增材制造产品的委托时，我们会建议用随炉试样做材料性能测试，因为产品本身往往不能破坏，而随炉试样能反映本次打印的实际性能。

力学测试我们建议覆盖不同打印方向，因为各向异性是这类工艺的固有特点，单方向的数据不完整。

对于多孔结构，清洁度和残留粉末的检查是我们建议的项目，因为这是外部看不到而又实际存在的风险。

如果你有增材制造的器械需要安排检测，可以把工艺、材料和产品结构发过来一起讨论方案，或者直接联系：132 4819 8029。能力范围见服务介绍，流程见检测流程，更多内容见知识库。

English version

Conclusion. Additive manufacturing is no longer novel in medical devices, appearing in custom implants, surgical guides, models and certain structural components. Its testing carries several distinctive difficulties. Consistency is hard to assure, since layer-by-layer building means small parameter variations affect the result, and each product may be different in custom production. Internal quality is hard to inspect, since lack of fusion, porosity, unremoved supports and trapped powder may exist inside where nothing is visible from outside. Properties are anisotropic, varying with build direction. And post-processing affects the outcome, with heat treatment, surface treatment and support removal each influencing final performance.

The consistency problem. The concept of a batch blurs, since conventional batch concepts do not transfer directly to custom production. Parameters are sensitive, with small changes in laser power, scan speed, layer thickness and powder condition all affecting results. Equipment condition matters through calibration and maintenance. Feedstock batches matter through powder variation and the use of recycled powder. And position matters, since parts at different positions within one build may differ. Positional variation within a build is distinctive to the process: temperature fields and gas flow within the build chamber are not uniform, and parts at the edge and centre may differ in properties.

Verifying consistency. Witness specimens built alongside each production build verify the material properties of that build. Positional coverage places witness specimens at different positions. Process parameter records capture the complete parameters of each build. Process monitoring observes the melt pool and layer images in real time. Periodic equipment verification uses reference parts to confirm machine condition. And powder management covers batches, recycling counts and property monitoring. Witness specimens are the core means of verifying consistency in this process, because the product itself often cannot be tested destructively, as with custom implants, while the witness specimen provides material property data for that build.

Detecting internal defects. Porosity from incomplete fusion affects strength and fatigue. Lack of fusion occurs between layers or tracks. Residual powder remains in internal channels and porous structures. Residual supports remain where internal supports were not fully removed. Detection methods centre on industrial computed tomography, which reveals internal structure; density measurement indirectly reflects porosity; and metallographic sectioning is destructive with high resolution. Industrial computed tomography is clearly valuable for these products, being non-destructive and revealing internal porosity, lack of fusion and trapped powder, none of which is visible externally.

Anisotropy. It arises because layer-by-layer building makes interlayer bonding differ from within-layer material. It appears as strength perpendicular to the layers generally lower than parallel. Its implication is that build orientation affects final performance. Testing must therefore sample and test in different directions. And design must relate loading direction to build direction. Testing should cover several directions, since single-direction data do not represent the material, and designers need direction-specific properties to orient the build sensibly.

Effects of post-processing. Support removal leaves a surface condition, possible residues and possible damage to the part. Heat treatment relieves internal stress and adjusts microstructure, affecting mechanical properties considerably. Surface treatment covers machining, blasting, polishing and chemical processing. Cleaning removes powder and processing residues. And sterilisation is more difficult for porous structures. Cleaning and sterilising porous structures is the difficulty: powder may remain deep within pores, and cleaning fluid and sterilant may struggle to reach, which needs particular verification for porous bone-ingrowth structures.

Surface characteristics. Roughness is generally higher with additive manufacturing. Partially melted particles adhere to the surface and may detach. Post-processed surfaces differ after machining or polishing. Porous structure surfaces require pore size, interconnectivity and surface condition. And cleanliness covers residues on surfaces and within pores. Adhering partially melted particles are a potential particle source: they are weakly bonded and may detach in service, which must be controlled for implantable products.

Organising verification. Process validation covers the combination of equipment, parameters and material. Build-level verification uses witness specimens for each build. First article verification inspects new designs comprehensively. Sampling inspection applies to volume production. And process monitoring captures real-time build data. Concentrate verification on process and monitoring, since inspecting every custom part fully is impractical and only a stable process can produce stable product.

How we handle it. For additively manufactured products we suggest material testing on witness specimens, since the product itself often cannot be destroyed while witness specimens reflect the actual properties of that build. We suggest mechanical testing across build directions, since anisotropy is inherent and single-direction data are incomplete. For porous structures we suggest inspecting cleanliness and residual powder, which is a real risk invisible from outside.

Send us the process, materials and product geometry and we will discuss the approach. Phone or WeChat: +86 132 4819 8029.