lecture 11 - IED Lecture 11 Class Goals For your Team...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: IED Lecture 11 Class Goals For your Team Project Establish Design Specifications Perform Competitive Benchmarking Generate Alternate System Concepts Experiments and Testing IED Establish Design Specifications What does the customer want Performance attributes Target values Team needs to agree on these Specs soon IED Perform Competitive Benchmarking What is the competition with Middle School Science demos ? What can you find on the web ? Do you still remember any of the demos from your past ? IED Generate Alternate System Concepts Brainstorming as a Team More "Dangerous" than as individuals for Mini Project Do not shut out a Team Member !! Write them all on a list IED Evaluating Competitive Concepts Decision Matrices useful But garbage in, garbage out One needs useful data In practice, would like to Test Then real data in Matrices IED Experiments and Testing Measurement and testing Experiments to develop design data, performance tests, quality tests, life tests Experiments to develop design data Validation of models Role of Prototypes Design of Experiments IED Experiments and Testing Life tests We don't do well in IED Application to IED projects What does apply, what does not apply to my project IED Measurement and Testing Four Categories of Tests Experiments to develop design data Performance validation testing Quality testing Reliability and Life testing IED Prototypes vs Experiments Prototype A first version of the final product Sometimes very useful to guide further development Particularly for user reactions Example Software IED Spiral Development Processes implement im le m p plan nt e plan 1st Prototype etc test 2nd Prototype test plan IED Prototypes vs Experiments Experiment Focus on what data we really need to know Best if results feed an analytical model The modern way IED Experiments to Develop Design Data Past "Build and bust" principle make a prototype, try it, watch how it fails, learn from the failure Still has a role in the early stage of development of a concept Sometime the best way to "get our hands around the problem" IED Experiments to Develop Design Data Current thinking Model and analyze based on known theory and data Think carefully about what additional information we need Design an experiment to IED Validate the "known theory" Get the additional information Might not look anything like a "prototype!" Validating Known Theory Why bother? After all, known theory is known, isn't it? What to do when you discover unexpected results. 1. Check if the theory applies 2. Check if the data is correct 3. Apply for a Nobel Prize IED Design of Experiments Example: Bread Making Process At least 11 parameters can be varied: amount of kneading temperature for rising time for rising number of kneed/rise cycles time for baking temperature for baking type of flour amount of water amount of yeast amount of sugar amount of salt IED Design of Experiments Example: Bread Making Process Seek to optimize some combination of features of the output bread Lightness Taste Texture IED Bread Making Process How many experiments would you need to conduct to determine the best combination ? What is OK and what is failure ? IED Bread Making Process You cannot take 3 months to determine optimum ! You need Design of Experiments ! Two readings in Library Reserve on this IED Why the Push to Reduce the Number of Test Runs? The Facts of Life Experiments cost $ - usually a lot Experiments take time - usually long Experiments take effort - usually a lot IED Summary - Experiments Experiments Get data needed for a particular design Use Design-of-Experiments to get to the "answer" (the model) with the l east effort, time and cost IED Helps to understand some of the science in reducing the design space to be explored Summary - Experiments Experiments Carry out the experiments to get the required data Product prototypes (usually most expensive) Bench top constructs (better) Numerical simulation (best, if it captures the physics) IED ...
View Full Document

This note was uploaded on 04/08/2008 for the course ENGR 2050 taught by Professor Morris during the Spring '08 term at Rensselaer Polytechnic Institute.

Ask a homework question - tutors are online