Quality Engineering
Quality Engineering
Quality Engineering
Begüm Okur
Selin Şirin
Teoman
Buse
Introduction
This project aims to simulate the production process and evaluate improvement processes using a
catapult system. The catapult serves as an ideal tool for modeling quality control and
improvement efforts in production processes. In this experiment, a catapult and a target board
were used to perform shots, and the results were analyzed. The accuracy and precision of the
catapult shots were evaluated using control charts, and process improvements were made.The
goal of the project is to ensure that the catapult shots consistently hit specific areas on the target
board, thereby maintaining and improving the production process. To achieve this goal, four
different experiments were conducted, each with varying configurations. The experiments were
The data collection process was conducted through four different experiments. In each
experiment, the catapult was used in different configurations to perform the shots. The data
collected indicated which area of the target board each shot landed on. The data were analyzed
using a scoring system as follows: Region 3 awarded 3 points (smallest circle), Region 2
awarded 2 points, Region 1 awarded 1 point, and Region 0 awarded 0 points (failed shot).
In the first experiment, the shots were performed from a distance of 30 cm with normal
tension, and a total of 100 shots were made. In the second experiment, the distance was closer,
with normal tension, and again a total of 100 shots were made. In the third experiment, the shots
were performed from the initial distance of 30 cm but with higher tension, and 100 shots were
conducted. Finally, in the fourth experiment, the distance was closer, and the tension was higher,
accuracy and precision of the data, each shot was observed, and video recordings were made.
These recordings will be used to verify the accuracy of the data if necessary. After collecting the
data and conducting the experiments, the data will be analyzed using control charts. Control
charts will help identify variations and areas for improvement in the process.
In this section, we analyze the data collected from the four experiments using control
charts by using python. The data analysis involves calculating the mean (X-bar) and range (R)
for each experiment, then constructing the control charts to monitor process performance. For
each experiment, we calculate the mean (X-bar) and range (R) for subgroups of size 5. Here are
the python code detailed calculations and control charts for each experiment:
import pandas as pd
subgroup_size = 5
subgroups = [experiment_data[i:i + subgroup_size] for i in range(0,
len(experiment_data), subgroup_size)]
A2 = 0.577
D3 = 0
D4 = 2.114
ucl_r = D4 * r_bar
lcl_r = D3 * r_bar
control_chart_data = pd.DataFrame({
'X-bar': x_bar,
'R': r,
plt.figure(figsize=(12, 6))
plt.xlabel('Subgroup')
plt.ylabel('X-bar')
plt.legend()
plt.grid(True)
plt.show()
plt.figure(figsize=(12, 6))
plt.xlabel('Subgroup')
plt.ylabel('Range')
plt.legend()
plt.grid(True)
plt.show()
The provided Python code is designed to analyze experimental data by generating control
charts, which help in understanding the stability and variability of the process. Initially, the code
imports necessary libraries, loads the data from an Excel file into a DataFrame, and extracts
relevant data for each experiment while cleaning any missing values. The core functionality of
the code is encapsulated in a function that divides the data into subgroups, calculates the mean
(X-bar) and range (R) for each subgroup, and computes control limits using statistical constants.
These control limits are then used to generate X-bar and R charts, which visually represent the
By calling this function for each of the four experiments, the code produces control charts
that allow for a comprehensive analysis of the process performance. These charts highlight
patterns, outliers, and overall process behavior, helping to identify the most stable and consistent
process. The analysis revealed that the first experiment demonstrated the highest stability and
consistency, making it the best-performing experiment among the four. This systematic approach
to data analysis and visualization is crucial for understanding and improving process quality.
First Experiment
The first experiment consists of 101 shots. The shots are grouped into subgroups of size
5, and the mean (X-bar) and range (R) for each subgroup are calculated. The control limits for
the X-bar and R charts are then computed using the constants 𝐴2, 𝐷3, and 𝐷4.
The X-bar chart shows that all subgroup means are within the control limits, indicating a
stable process. The process average (center line) is consistent with minimal variation.
The range chart indicates that the process variability is within control limits, suggesting
calculate the means and ranges for subgroups of size 5 and to compute the control limits.
The X-bar chart shows that most of the subgroup means are within the control limits, but
some points are near the control limits, indicating potential shifts in the process mean.
The range chart indicates that the process variability is within control limits, but there is
The third experiment also consists of 51 shots. The same calculations are performed for
The X-bar chart shows that some points are near the control limits, indicating increased
variability in the process mean. The process is still within control limits but shows signs of
instability.
The range chart indicates higher variability compared to the first and second experiments,
Forth Experiment
The fourth experiment also consists of 51 shots. The calculations for the X-bar and R
The X-bar chart shows that most subgroup means are within control limits, but there are
signs of shifts and trends, indicating that changes in distance and tension have affected the
process.
The range chart indicates that the process variability is within control limits but shows
Summary
Based on the analysis of control charts and process performance, the first experiment is
selected as the best experiment due to its stable and consistent results. The first experiment's
shots are more accurate and exhibit less variability, making it the most reliable process among