In the design and manufacturing of any product, there will always be failures, and whether anticipated or not, you can bet that they will come when you least expect them. Determining the cause and who carries what part of the responsibility can be a sensitive, yet critical matter. You need to be confident that if design changes are warranted, they are effective in eradicating the failure.Here, let’s focus on the design side and utilize the following approach.
Step 1: Define the problem and its scope (evidence is key)
- How long has the problem been happening (years, months, days, or once) and how many have failed (one part, one lot size, or one gearbox model)? On the design side, this will help you understand how sensitive the product may be to the application. On the business side, this will help you determine a budget to fund the exposure.
- Have a thorough discussion with your customer to understand exactly what the circumstances were when the failure occurred and, specifically, what was the problem. Your customer may not completely understand how their product is used. If this is the case, have someone on your technical team visit the end-user site and conduct interviews. Speaking with the operator of the product can uncover valuable information that they felt was insignificant.
Step 2: Gather the data
- Collect all of the available data such as hours, loads, application, population size, design changes, supplier or manufacturing processing changes, customer changes, and how many items have shipped. Utilize statistical and reliability tools such as Weibull analysis, MTBF (mean time between failures), and MTTF (mean time to failure) to help quantify the actual life and uptime. Do the results fall into gear premature failure or expected design life?
- Ensure that what you designed and delivered was capable of meeting expectations. Check your design file and confirm that the actual duty cycle was known, your design team calculated performance to the expected duty cycle, and that an appropriate safety factor was achieved. Specifically, were the end user’s duty cycle, system inputs, system outputs, and the environment thoroughly called out? Did the customer acknowledge that the design met the requirements? Did your design team understand and apply applicable design protocol? Was the appropriate safety factor for the intended environment and use applied? What were the system’s dynamics, and were they understood? (It is not uncommon for a component or subassembly to prematurely fail because of an element outside of your design control. Parts of the assembly, the power input, mounts, or isolators could contribute to or initiate a system excitation.)
- Validate that what you manufactured met your own design specifications. Did the design team take into consideration the impact of manufacturing’s variability? Do your manufacturing records and test results confirm compliance?
- Get all the related parts and, if possible, any contributing components like shafts, couplers, clutches, and PTOs. Find their design and performance-specific characteristics.
- Preserve the as-received condition of the parts. Conduct a thorough visual inspection prior to any disassembly and capture what you find with photographs and a written summary. Is there any evidence of misuse, overloading, corrosive elements, or parts exhibiting yield? Were the parts and assemblies maintained and lubricated properly? Do the failed parts indicate brittle or fatigue failure? (Take high-resolution digital photographs of any failed surfaces.) A quick refresher in your metallurgical references can help you here.
- Inspect the mating gears, shafts, housings, bearings, and add-on components. Collect samples of the oil or grease and have them analyzed. Look for contaminants and debris that may help identify contributors to failure.
- Be sure to employ outside services if necessary. Results from providers of specialized services such as FEA (finite element analysis), test labs (to provide fatigue, destructive, metallurgical, chemical, or nondestructive testing), and system dynamic response analysis in the field can be invaluable.
Step 3: Identify the root cause
- Be sure that the team working on the problem has a balanced skill set.
- Generate many contributory probabilities and scenarios. Implement brainstorming, cause and effect diagrams, and pareto charts.
- Review your data; challenge any changes in process, design, or use.
- Facilisis in pretium nisl aliquet
- Develop a short list of probable causes, and select the most plausible. At this point, you should have a clear understanding if the failure was related to an inadequate design or attributable to other factors.
Step 4: Implement changes
- Once you have a likely cause, then you can redesign the part, system, or process to specifically address that concern.
- Make changes that will outpace the variability in use. Aim for roughly a 20- to 25-percent improvement in the criteria.
- Test the solution in a real-world application if possible before you turn the production switch back on.
There will always be failures, but delivering the right solution the first time is critical for your customers and your business.
