How can you reduce risks and uncertainties in mechanical design?

After all these, what I think is basic and most important is to thoroughly understand the working environments and the service conditions as these inform the input parameters to design. When design requirements are clear, and management of change process is established, design risks are minimized. Design is an iterative process and user experience should be taken to cognizance at every time. No design is perfect. Designs are improved with time using technology. However, as a designer, know that if anything must fail, it should fail safely, hence the need for constructability reviews, HAZID, and HAZOP studies.

I believe the first step for defining the design general features needed is site/environment inspection. Through detailed inspection of the environment and the site, We shall have the chance to know some parameters like: Temperature, humidity, installation position, dimensions and tolerances, accessibility and maintainability. This detailed analysis and researching is critical as it will result in high effectiveness and reliability of the product as well as savings in time and cost. (CAPEX & OPEX)

In many aspects of a mechanical design, risks and uncertainties help to shape the eventual solution. As such, in the manner of agile design approach, a quick-iterative-functional (QIF) prototype should be prioritise for testing in the real environment. For projects where the real enviornment might be difficult to reach, for example satellite or spacecraft, the approach can take the form of a MVP (minimum viable product) to be tested in simulated environment such as hot/cold plate, thermal and/or vacuum chamber, and vibration platform.

It is important to have a cross functional team while carrying out DFMEA analysis as all risks need to be thought upon and evaluated for criticality and occurence.

Risk management tools are great during the project phases before industrialization. Once your product reaches sampling phase then you see if all your planning was put to good use. Analysis of these first samples is crucial to improving your design and mitigate future risks. When working with hundreds of components which need to fit together and engineers whom each think their design will work 100% then your once sure fire solution becomes uncertain.

DOEs are crucial in all stages of product development, especially in the design discovery phase of first generation products in which the physical phenomena may not be known fully and/or the critical design parameters are not determined or data driven for the design outputs in interest. The sensitivity analysis achieved by DOEs will enable the designer to focus on the key design parameters to shape the main architecture and make the most impactful decisions to guide the design and manufacturing from the very beginning.

Artificial Intelligence can handle all of theses, starting from Generative AI--> CAD drawing --> 3D Modeling and test --> Modify up to get the optimum product.

Topology optimisation techniques are helpful in getting an intuition on the right shape early in the design process. This saves the time of design engineers and computational efforts of simulation engineers as it avoids iterative efforts.

DOE and Optimization methods can help to bridge the gaps between the design variables and the responses which can help the design to reach the optimal conditions.

Utilizing design of experiments and optimization is vital for effective mechanical design. Central Composite Designs enhance factorial designs, revealing deeper insights into design spaces. Lattice Designs efficiently analyze multiple factors, while Multi-Objective Optimization assesses various criteria, using Pareto analyses for trade-offs. Evolutionary Algorithms, inspired by nature, explore optimal solutions, complemented by Metaheuristic Techniques. Surrogate Modeling accelerates optimization when direct methods are resource-heavy. Machine Learning, especially neural networks, refines designs by leveraging large datasets. Together, these techniques drive optimal design outcomes.

I think a robust design can be made by understanding how your product will be manufactured. The design parameters and tolerances change according to the processes. For example, the tolerances for a casted product will be different from that of a machined product. One can try to use standard products to reduce manufacturing time, but the tolerances of such standard products need to be considered if they will be connecting with other parts in the design.

Robust design principles are essential for reliable mechanical design outcomes. Central to this is building resilience against unforeseen factors, ensuring not just quality but consistent performance under diverse conditions. Addressing external disturbances, like unexpected operations or user behaviors, is crucial. Implementing tolerance analysis guarantees parts operate effectively in real-world situations. Using feedback loops refines design based on real data, and system simulations predict overall effects. This comprehensive approach ensures designs are both optimal and consistently effective across different scenarios.

To reduce risks and uncertainties in mechanical design, applying robust design principles is key. This involves designing components and systems to be insensitive to variations, including manufacturing tolerances, material properties, and environmental conditions. Techniques such as Taguchi methods, sensitivity analysis, and simulation-based design optimization are employed to achieve designs that perform reliably under a wide range of conditions, thus minimizing the risk of failure and ensuring product durability and reliability.

I think that will be also covered by some AI solutions. Generative AI-->CAD-->CNC-->Asset-->Test-->Return to Generative AI to modify.

Avoiding notches on the designed part as much as possible. Avoiding sharp corners and bends can go a long way in improving the life of components. Keeping the design simple can help in making the design robust.

Prototype in 90's and before. Was take many steps to be life and design engineers work hard to reduce the cost of re-mafufactoring it. Now you can design prototype and print it using 3D printer with any required size.In 30 mintues you can testing, inspection, simulation, modeling, and analysis to verify and validate your design.

While we validate our parts, we need to ensure that the loading scenario and frequency is replicated on the test bench. Effects of unreliability can be evaluated by calculating B10 life using different slope factors. Validation processes can be expedited by using accelerated test methods.

In my experience and in my current job profile,We use the Reputed software for mechanical design which are based on ASME code.hence before using any software we have to validate the software with respect to any code or standard in my case ASME code. After validation only we can find out the bugs and error in software. Sometime it will huge impact on design which is a directly related with material cost. We cannot blindly trust on software when it release to market after any updation.

As an AI each design operation saved in the database for further action and to help improve design operation. Some companies allow AI to think out the box to design new shapes or surfaces. That help manufactors produce a product sheap, save and ofcurse beutiful.

Lessons learned from the Aerospace Industry may provide insights for other industries requiring safety assurance and risk management. The grouping of parts based on the similarity of risk assessment of application, design, and processes, which we refer to as PRAGO - Part Risk Assessment Grouping and Oversight, takes into consideration process maturity evaluation and risks as well as Form, Fit and Function. Tools for risk assessment may include but not be limited to: Design and Process Failure Modes and Effects Criticality Analysis (DFMEA, PFMEA, FMECA) and Operating Environment Assessments (duty cycles, sand, pollution, salinity, variable loads, Foreign Object Damage, etc.)

Attention should also be given to common mode failures which can be missed through risk evaluation strategies. Lapses in maintenance procedures should be avoided.

The level of control and oversight must be balanced with delivery and cost considerations. Well-conducted risk assessments help establish and communicate the necessity for suitable fixed costs of compliance and oversight. The level of effort to conduct such risk assessments can be mitigated by grouping and assessing 'similar to' parts based on application, design and processes (PRAGO - Part Risk Assessment Grouping and Oversight). Special processes (output cannot be verified by subsequent monitoring or measurement) may require process approvals, auditing and approved personnel. Higher risk applications, parts and processes may require part specific, fixed or frozen practices for manufacture, assembly and repair.

Check the tolerance stack ups, especially if many tolerances affect an important feature of your design. Make sure that the worst case scenario leaves enough material to survive your loads and operating environment.