An Innovative Predictive Model of Catapult Performance

Overview

Objectives/Goals We examine the consequences of using mathematical and statistical modeling to characterize the functional performance of a novel catapult design. Real-world projects are complex and runs are costly. Our project explores fractional factorial experimental design to characterize a mechanisms performance. The design allows for testing of multiple factors at an acceptable number of runs. Methods/Materials The novel mechanism is a catapult with a range of up to 4 m. We will: 1. build a novel catapult that produces reliable, predictive functional performance; 2. run a fractional factorial design of size 2^(5-1) (resolution IV); 3. create mathematical and statistical models to predict functional performance. We will compare the catapult performance of a theoretical model to the predictive capability of fitted statistical model. The catapult will be subjected to tests using a semi-randomized experimental design with 20 runs (2^(5-1) + 4 centerpoints), and five three-level factors. The response effects for factor settings will be assessed by comparing responses at high- and low-level settings. The most significant effects for given factor settings will be used to construct a multiple regression model. The predictive capacity of the fitted regression model will then be tested. The theoretical model will be compared to the statistical model. Results Projectile launched towards 1 m, 2 m, 3 m had average values of 1.09 m, 2.01 m, and 3.08 m, respectively. Shot distributions were approximately normal distributed in agreement with our models. We verified the goal of building a novel, precise, mechanical apparatus, since all factor effects were clearly distinguishable from random noise. Additional predictive tests of the fitted model exhibited performance at a 95% confidence level. The catapult was not equally capable at all targets; however, resulting impacts fell within the hypothesized 0.15 m range of the intended target. This is a remarkable level of predictive capability. Conclusions/Discussion The goals of the project were fully attained. More advanced studies of catapult performance could be done: 1. we could investigate quadratic effects in the experimental design; and 2. we could use advanced multiple regression analysis, such iterative multiple regression analysis (IMRA).

Awards (1)

• Second Place

Competition history

• CSEF 2011 Applied Mechanics & Structures · Entry S0314

Resources

• CSEF project summary ↗

Related projects

Closest projects by meaning, across every fair and year in the corpus.

Browse more like this

• All Engineering projects
• CSEF Engineering projects
• Every CSEF 2011 project
• CSEF Second Place winners