ME453_Lab1_FA23

pdf

School

University of Illinois, Urbana Champaign *

*We aren’t endorsed by this school

Course

453

Subject

Mechanical Engineering

Date

Dec 6, 2023

Type

pdf

Pages

4

Uploaded by DukeMeerkat3564

Report
Page 1 of 4 Lab 1: Getting Ready for Manufacturing Quality Control Assigned: September 6, 2023 Due: September 27, 2023 Lab objectives (1) Perform ultrasonic metal welding experiments and collect online sensing data for quality monitoring purpose. (2) Perform preliminary data analysis including visualization, exploration, frequency analysis, and hypothesis testing. (3) Practice data science thinking in a real-world manufacturing/engineering setting. Introduction Ultrasonic metal welding is a solid-state joining process that is well suited for battery tab joining in electric vehicle manufacturing. A typical ultrasonic welding system is displayed in the figure below. During each welding cycle, the transducer transforms electrical energy into high- frequency mechanical vibration. This mechanical vibration is transferred to a welding tip through an acoustically tuned horn. A metallurgical bonding between thin metal sheets clamped under pressure is created with oscillating shears generated by high-frequency vibration. Ultrasonic metal welding can produce good quality welds when the welding conditions are optimal. However, an abnormal process condition can cause poor welds even with optimal parameter settings. For example, workpiece and tool surface conditions strongly affect weld quality. Vanishing oil, stamping fluid, cutting fluid, and other oil-based fluids are reported in the assembly line of battery packs as possible sources of contamination. The presence of oil contaminations will lead to poor weld quality. Unfortunately, such abnormal process conditions may not be easily detected before a welding cycle. Online process monitoring with sensing signals is a common approach in ultrasonic welding of automotive lithium-ion batteries. The potential quality of the weld could be determined by the features extracted from the online process signals. In this lab, we will develop a prototype online monitoring algorithm to detect abnormal
Page 2 of 4 process conditions utilizing features generated from sensing signals. The abnormal process conditions will be simulated by applying Tap Magic ProTap Cutting Fluid at the interface between the workpieces. Besides, the effect of the workpiece dimension is also studied in this experiment. Experimental settings Use copper sheets for both top and bottom layers. Dimensions of sheets: Fixed width and 4 levels of length The thickness of sheets: 0.008” Each group will work on a different dimension setting. 30 samples will be welded at three different levels of surface condition: o 10 at level 0 (clean with ethanol-moistened cleaning wipes) o 10 at level 1 (Wipe the welding area of one sheet with oil-moistened cotton swabs) o 10 at level 2 (Wipe the welding area of both sheets with oil-moistened cotton swabs ) Before you start (1) Open “DataCollection.m” in MATLAB and run the “Initialize” section to initiate the data acquisition system. Define “defCn” as the current cycle number. (2) Set the sampling rate at 20kHz. (3) Run a test weld to check whether the code works fine. After each weld, the code should display plots and write a csv file in the same location. The csv file will be named by the cycle number (5 digits) automatically. For example, 19677.csv (4) Now you are ready to collect the sensing data. Data acquisition and experiment procedure (1) Make sure the welder is in “Time” mode and the weld time is set to 1 second. Make sure the maximum power is set to be around 1900 W. Otherwise, the controller will give a warning. For each weld, set the Pressure as 50 psi and the Wavelength as 40 microns. (2) Prepare the copper sheets with the correct surface conditions (clean, contaminated, polished). (3) Write the cycle number on the sample as displayed on the controller screen. (4) Record the cycle number in the excel sheet. (5) Run the “Data collection” section in MATLAB by pressing “Shift + Enter”. (6) Confirm the cycle number in the pop-out window. (7) Place the copper sheets between the horn and the anvil and start welding. Carefully align the samples and make sure the welding tip is above the center of the copper sheets. (8) Make sure you get figures displayed in MATLAB and a csv file saved in the current folder. (9) Each csv file has 6 columns and data for a duration of 2 seconds.
Page 3 of 4 a. 1st column = time (sec) b. 2nd column = Acoustic Emission (AE) sensor signal c. 3rd column = Displacement signal from the actuator d. 4th column = Power signal from the controller e. 5th column = Sound signal from the microphone f. 6th column = Clamping pressure from the pressure regulator (10) Run the next experiment with the same procedure above until all 30 samples are completed. Once you are done Copy all the csv files to a folder named by today’s date and workpiece dimnesions. For example, “230904 _ 2 inch”. The data will be shared through Box. Data analysis and lab report A Python template is provided to assist you with the data analysis. You will work on a combined dataset with all groups’ experiments. Note that data collection is worth 10 points and the following analysis tasks are worth 90 points in total. (1) Influence of tool conditions. i. Use data collected from experiments with level 0 samples to visualize the influence of workpiece dimensions. Specifically, plot the power signals with the same dimension in one figure. You will need to generate 4 plots for 4 workpiece dimensions. (10 points) ii. What differences can you tell from the plots? No quantitative measures are needed. (5 points) (2) Influence of surface contamination. i. Use data collected from experiments with “2 inch long workpiece” to visualize the influence of surface contamination. Specifically, plot the power signal with the same surface condition in one figure. You will need to generate 3 plots for 3 contamination levels. (10 points) ii. What differences can you tell from the plots? No quantitative measures are needed. (5 points) (3) Frequency analysis. i. Perform Fast Fourier Transform (FFT) for all acoustic emission and microphone signals. (10 points) ii. Record the peak frequency around 20 kHz ( ࠵? ) and the corresponding power value ( ࠵? ). (10 points) iii. Calculate the mean and standard deviation of ࠵? and ࠵? for each combination of workpiece dimension and surface contamination level. You will need to calculate 4 × 3 = 12 numbers for both ࠵? and ࠵? . (10 points)
Your preview ends here
Eager to read complete document? Join bartleby learn and gain access to the full version
  • Access to all documents
  • Unlimited textbook solutions
  • 24/7 expert homework help
Page 4 of 4 iv. Plot box plots for ࠵? and ࠵? against the condition combinations. (10 points) (4) Statistical hypothesis testing. i. Using data from level 0 samples with varied workpiece dimensions, carry out t- test to examine if ࠵? and ࠵? are significantly different between workpiece dimensions. (10 points) ii. Using data from samples produced with the “2 inch long workpiece” but different sample surface conditions, carry out t-test to examine if ࠵? and ࠵? are significantly different between sample surface conditions. (10 points)