Short: This data is based on an DSD with a couple of repeats and different extraction points for cross sections of the weld to model the whole weld depth curve over the weld which is not uniform over the weld length. You can use it to show FDE and FDOE in the context of laser welding but also when a uniform curve is needed for a customer.
Background Story
Many processes demand a uniform quality metric over a specific time or length to ensure uniform quality. This example shows some real data of a laser beam welding process of dissimilar metals where the weld depth in the lower sheet determines the mixture of the metals and ultimately hot cracking in the weld. To ensure a crack free welds the weld depth should not exceed roughly 250 µm over the whole weld.
The data is based on a DSD with 6 factors and 18 runs. Every run was repeated 5 times and the weld depth at 5 different cross section points over the weld length was measured. The data can be analysed with FDE and FDOE.
Some key findings you can communicate:
- Stronger weld depth uniformity with rising heat input (high beam power; low welding speed)
- Curvilinear behaviour of Focal Position (lower penetration when defocusing the beam)
- possible Interaction between Focal Position and Angular Position
- Optimize for target function
- Use the profiler to find a laser beam power curve to counter the remaining weld depth uniformity
|
Job Title
|
Process Development Engineer; Process Engineer; Research Engineer
|
|
Industry
|
Electronics Industry; Aerospace Industry; Medical Industry; Automotive Industry
|
|
Critical Business Issue
|
Need to develop highly automated and fast laser beam welding processes with defect free welds.
|
|
Problems/Reasons
|
1. Many factors influence the weld metal dimensions and weld quality
2. Metrics like weld depth can change over the length of the weld and can influence the weld quality as well as the weld strength
3. Need to optimize the process for uniform quality to avoid customer complains and time consuming rework
|
|
Specific Use Cases
|
Need to build a process model which incorporates all process variables as well as the changing quality metric over the weld length
|
|
JMP Capabilities
|
DoE: Definitive Screening Designs (DSDs), Custom Design; Functional Data Analysis and FDOE
|
|
User Wins
|
1. Far better process understanding and easy optimization according to customer needs (e.g. saving energy with low power or short process time with high welding speed)
2. Optimize for uniform weld depth over the weld length as well es information about possible power curves to counter weld uniformities
3. Great way to communicate results with customers using the profiler
4. Higher chances of promotion/standing out for successful and timely project closure
|
|
Business Results
|
Better process understanding, less rework due to better quality, informed decision making, better customer relationship, less stressed employees.
|
