Integrating Experimental Design and Reliability for Multiple Stage Manufacturing of Multi-Scale Devices

这笔拨款为开发和分析基于统计的实验设计提供了资金，特别适合于多阶段制造，并将这些设计与新产品开发的可靠性建模相结合。提出了多级分数因子分割图（MSFFSP）设计的概念，以适应复杂的制造过程，其中需要对实验运行的随机化进行限制。具体而言，将推导出MSFFSP设计的一般线性模型、设计目录和方差估计。这项工作将应用于包括凝胶铸造光刻在内的多阶段工艺，以制造用于微创手术的多功能钳子-剪刀（FS）器械。新的光刻工艺使用纳米颗粒赋予FS仪器具有纳米和微观尺度的特征。MSFFSP设计还将应用于光学显微镜的测量系统分析，该显微镜用于在FS设备上收集数据。结合FS设备的开发，将识别、列举故障模式，并将其转换为可靠性数据。加速测试的概念将用于缩短获取这些数据的时间，并将与MSFFSP设计集成。为了正确地分析可靠性数据，将对MSFFSP分析进行修改，可靠性数据与过程和产品数据具有不同的统计特征。如果成功的话，这项研究的结果将导致在保持设计效率的同时，在多阶段的复杂制造过程中进行实验所需的试运行次数和设置的减少。使用这些设计，并将其与可靠性测试集成到微纳米级制造研究的建模和生产中，将通过加快研发周期、延长实验预算，最终帮助创造更可靠、更健壮、性能更好的小规模产品，从而产生战略优势。所提出的工作不仅将推进实验设计领域，而且还将扩大质量方法的潜力，以扩展到其他多阶段制造过程。

This grant provides funding to develop and analyze statistically-based experimental designs especially suited for multiple stage manufacturing, and to integrate these designs with reliability modeling for new product development. The concept of a multiple stage fractional factorial split plot (MSFFSP) design is proposed to accommodate complex manufacturing processes where restrictions on the randomization of experimental runs are required. Specifically, a general linear model, design catalogs, and variance estimates of the MSFFSP design will be derived. This work will be applied to a multiple stage process involving gel-casting lithography to manufacture a multifunctional forceps-scissors (FS) instrument for minimally invasive surgery. The novel lithography process uses nanoparticulates to impart the FS instrument with both nano- and micro- scale features. The MSFFSP design will also be applied in the measurement system analysis of an optical microscope, which is used to collect data on the FS device. In conjunction with the FS device development, failure modes will be identified, enumerated, and transformed into reliability data. The concept of accelerated testing will be used to shorten the time to acquire this data and will be integrated with MSFFSP design. The MSFFSP analysis will be modified in order to properly analyze the reliability data, which has different statistical characteristics than process and product data. If successful, the results of this research will lead to a reduction in the number of trial runs and settings required for experimentation in complex manufacturing processes with multiple stages while maintaining design efficiency. Using these designs and integrating them with reliability testing into modeling and production for micro- and nano-scale manufacturing research will yield strategic advantages by speeding the research and development cycle, stretching the experimental budget, and ultimately helping to create more reliable, robust, and better performing small-scale products. The proposed work will not only advance the field of experimental design, but also amplify the potential of quality methodologies to be extended for other multiple stage manufacturing processes.