Design for manufacturing and assembly (DFMA) units

When a system for a special use is conceived, with limited precision, and unlimited durability (eg a satellite), we are strongly constrained by the particular technological challenges involved, by the time available, and by a budget (CAIV: cost as an independent variable) 

The perverse loop is that restricting ourselves even more, thinking of manufacturability and mutability, complicates the problem devilishly, and very often we give up even knowing that it is precisely this tactic that can help us relax our bonds.

Design For Assembly

Tensions that are tense are fatal to achieve the generation of ideas of value that overcome technological challenges and achieve an innovative result with a reliability of exploitation 6Sigma that prolongs his life. But, on the other hand, excessively weak ligatures make the well-known “influence-cost incursion” curve be pronounced even more, with its fatal consequences in terms and costs out of budget.

To all this is added the problem that compliance with the basic law “the best design is the simplest that works”, is much more difficult to verify when we are conceiving complex products of unique characteristics not experienced before, and with specifications frequently changing during your lead-time until delivery.

Design for Manufacturing

Then, it is already understood that, depending on which approach we apply DFMA, we can achieve a beneficial effect (in functional value, reliability, term and cost) or catastrophic. And, of course - we would all do the same, and well thought out -, in the face of the situation, we are left with "scrupulously comply with the product, and with the process do what we can Because we have learned well from the "catastrophic" side of reliability, costs and deadlines with two recent examples of the aerospace sector; the F-35 “lighting” of Lockheed, and the B-787 of Boeing. Everything has gone wrong, and that they were not "unitary projects"!

With what features should we apply DFMA in these cases?

To answer this we must refer to two previous articles entitled “Brief note on how to generate ideas of value for its generalized flow in design and industrialization? And “Extreme agility and quality in the design and industrialization of complex products with high content in technological innovation”, which define, respectively, where and how ideas of value (functionality-reliability) are generated, and how their flow is organized.

Durability and reliability with the basic specifications and the technological challenges involved.

Specifications "augmented" on their own initiative to surprise the customer.

Flexibility when changing specifications "on the fly", without getting into "traps" for tools or special, specific or rigid tooling.

Tight terms, typical object of the traditional synchronization Lean.

CAIV, costs as the “objective variable”, also the object of traditional Lean, a method whose strict application would be myopia and a disaster with certainty.

(As it is seen that they are only in fourth and fifth priority, the traditional Lean, VSM, PDVSM, etc. are not the key to modern design, by the way, very badly called “Lean Design”).

We are not going to repeat here the super-known DFMA application principles, generic as all engineering assistance techniques, and too focused on the world, much easier, of the repetitive product and / or medium-long series. I will only emphasize those features that I see most applicable to this context with the aforementioned priorities, 1 to 5.

They are not feasible due to high risk against priorities 1, 2 and 3:

• The elimination of "non-essential parts and components" (which only hold, are static, 
• Etc. because it is better to ensure the uncertainty of an innovative and complex design.
• Standardize materials and eliminate diversity, for the same reason.
• Avoid difficult components, in the sense of manipulation, fragility, etc.

They are moderately feasible:

• Use some (not many) multifunctional components in different parts of the project. Even screws, rivets, etc. because in these cases, of "commercial hardware material", they have nothing, they are special and expensive.
• Try an assembly as “telescopic as possible” on the Z axis, taking advantage of the forces to be used only to “stop” gravity (which is much more precise and safe), and to make the final adjustment.
• The reduction of tooling and fixturing specific for positioning, assembly and tests.
• The application of economic fabricability criteria (orthodox application of DFMA) to the components themselves. Typically this occurs in "fiber placement" of carbon fiber composites, in vertical turning, and in high speed machining by prismatic machines or by hexapods, etc.

They are, in general, the most feasible:

• Avoid fasteners and fastening points, which are the focus of definability and cost of assembly and manufacturing.
• Organize the assembly to avoid the "stack" (undue accumulation with bias) of tolerances, which would lead to strong adjustments, residual stresses and again away from operational reliability> 6Sigma. Also measure it in real time, and act from an intelligent SW (such as Promina, in the case of Sister plant), giving instructions for the alteration of requirements of tolerances or changes of sequences in the later phases of the assembly.
• Avoid "blind" assemblies in which you have to act without directly seeing the object, and are not accessible to a hexapod-type robot - ideal for its many degrees of freedom.
• Give clear accessibility to maintenance or tests with the minimum possible disassembly of any kind, including partial ones.
• Locate “fasteners” in the assembly components themselves, as well as “centering and safe positioning” points.
• Design modular subsets that can be tested at their own level, avoiding the uncertainty of the final tests, and saving specific tooling in addition to a lot of time. A case of particular interest is the centers of gravity.

In conclusion, I would say that the application of DFMA in the industrialization of products with high technological added value (innovative, that overcome complex technical challenges and unusual functionality, and that must be reliable> 6Sigma) must be restricted precisely with the criteria that I have explained in this article. The opposite has the inadmissible risk of ruining the project, very early or only somewhat later due to lack of reliability.