Reliability Improvement of Lithium Cells Using Laser Welding Process with Design of Experiments

Manufacturing operations introduce unreliability into hardware that is not ordinarily accounted for by reliability design engineering efforts. Inspections and test procedures normally interwoven into fabrication processes are imperfect, and allow defects to escape which later result in field failures. Therefore, if the reliability that is designed and developed into an equipment/system is to be achieved, efforts must be applied during production to insure that reliability is built into the hardware. There are various ways to improve the reliability of a product. These include: Simplification Stress reduction/strength enhancement Design Improvement Using higher quality components Environmental Stress Screening before shipment Process Improvements, etc. This paper concentrates on 'Manufacturing Process Improvement' effort through the use of design of experiments, (DOE). Hence, improved levels of reliability can be achieved. Most applications optimize the mean value of a response variable. Parameter design experiments are aimed at reducing variability caused by manufacturing variations. In industrial processes, controlling variability is usually much harder than controlling the mean value. And variability is often a root cause of manufacturing problems. Optimizing the process of welding the lithium cells hermetically, by laser, have been focused in this work. Cell cans were approximately 1.5 inch diameter, nickel-plated, cold rolled steel with the top welded on around a 0.040 inch lip. The important characteristic measured in this work is weld penetration depth. The key steps in this project are: 1. Identify important process parameters / factors, (controllable and un-controllable / noise), and their potential working levels that can be manipulated. 2. Design and perform 'Fractional Factorial Experiments' on the process. 3. Analyze the resulting data to determine the optimum operating levels of the factors such that the product's functional characteristic is optimized and has minimal sensitivity to noise.