1 / 37

ChE 551 Lecture 04

ChE 551 Lecture 04. Statistical Tests Of Rate Equations. Last Time Considered Paramecium Example. r 2 does not indicate goodness of fit. Today: Statistical Analysis Of Rate Data. Can we do a calculation to tell if one model fits the data better than another model?

Télécharger la présentation

ChE 551 Lecture 04

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. ChE 551 Lecture 04 Statistical Tests Of Rate Equations

  2. Last Time Considered Paramecium Example r2 does not indicate goodness of fit

  3. Today: Statistical Analysis Of Rate Data • Can we do a calculation to tell if one model fits the data better than another model? • Is the result statistically significant?

  4. Method: Calculate A Variance Usually model with the lowest variance works best! (3.B.1) substituting in equations (3.A.7) yields (3.B.2)

  5. Limitations Of Using Variance To Assess Which Model Fits Best • Assumes error in data • Follows a “2 distribution” (i.e. error is random) • Usually good assumption in direct rate data • It is not good to assume 1/rate follows 2 distribution, so one needs to be careful about linearizing data.

  6. For Our Example (3.B.3)

  7. For Our ExampleContinued Eadie-Hofstee: (3.B.4) while for the Lineweaver-Burk Plot: (3.B.5) The non-linear least squares fit the data best.

  8. Next: Using An F-Test To Tell If the Difference Is Statistically Significant Method: Compute Finverse, given by (3.B.6) If Finverse is large enough, the model is statistically better.

  9. Statistics: Gives A Value of Finverse That Is “Large Enough” nf=number of data points - parameters in the model (3.B.8) To read the table, if nf=4, you need Finverse to be at least 15.98 to be 99% sure that the better model really is better. There will still be 1% chance that the differences caused by random errors in the data

  10. Assumptions In Using the Values Of F In Table 3.B.2 • Models are independent (non-nested) • 2 distribution of errors Not mathematically rigorous in our example since models not independent! (Gives small error in Finverse)

  11. Fdist Gives The Probability That A Given Model Is Better % confidence=1-FDIST (Finverse, nf for better model, nf for worse model) (3.B.9) Not mathematically rigorous, but close.

  12. Example: Is The Non-Linear Least Squares Better Than LineWeaver-Burke Variance Lineweaver-Burke = 321 Variance non-linear = 185 nf=30 I used Excel to calculate 1-FDIST (1.92, 30, 30)=0.96 96% sure non-linear least squares fits better 4% chance difference due to noise in data.

  13. Another Example: Comparing Two Models Previously fit data to (3.A.1) Does the following work better? Is the difference statistically significant?

  14. The Spreadsheet Is The Same As In Problem 3.A:

  15. F Test To Determine Which Model Is Better V3.A.1 the variance of equation 3.A.1 is V3.C.1 the variance of equation 3.C.1 is The ratio of variance is

  16. Calculate Probability Second Model Is Better From FDIST probability=1-FDIST (1.07,30,30)=0.58. 58% chance second model is better 42% probability first model is better Note: Not rigorous number

  17. Next: Multiple Variable Analysis • Rates of reaction usually strongly effected by many variables • Temperature: concentration, solvents, inpurities, catalysts, …… • So far only consider one variable: Concentration

  18. Example: Develop A Rate Equation For The Growth Of Grass Variables • Sunlight • Rain • Amount of grass seed • Number of birds and insects • Fertilizer • Soil type • Soil bacteria How do we proceed to measure a rate?

  19. Usual Technique: Initial Rate Method • Start with multiple parallel reactors • Fill each with a different concentration • Let reaction go & measure conversion vs time • Get rate from slope extrapolated to zero

  20. If We Have Several Variables, What Do We Measure? General approach • Take some preliminary data to determine what variables are important • Usually requires multiple iterations • Take more detailed measurements on the variables that are most important

  21. Design Of Experiments To Determine Which Variables Are Important • 2n designs • Pick two values of each of the variables • Look at two possibilities for each variable • Do experiments for all combinations • Do analysis to decide which variables are important

  22. Example: How Does Temperature And Concentration Affect Selectivity Of A Reaction • Pick two values of each variable Temperature + = higher temperature Temperature - = lower temperature Concentration + = higher concentration Concentration - = lower concentration • Look at all possibilities

  23. Table of All Possibilities

  24. How Do We Analyze The Data? • Look at the deviation from the mean • Calculate row averages

  25. For Our Example, Mean=45%

  26. Calculate Row Averages

  27. First Conclusion • Want temperature to be low • Cannot tell about concentration

  28. Calculate Row Averages

  29. Is It True That We Do Not Care About Concentration? Answer no: If the temperature is low, can improve conversion by keeping the concentration high – it is just that the opposite effect occurs when the temperature is high

  30. Lets Examine The Effect Of TC (Simultaneous Variation of T+C) Want T – and TC -

  31. Can Extend Process To Several Variables Gives too many runs

  32. Software To Help • Concept: we usually want to fit the data to a simple function: Response=C1+C2A+C3B+… Only need enough runs to fit constants accurately

  33. Echip Software Example

  34. Software Setup

  35. Number of Runs Substantially Reduced • 4 variables, 4 values with 3 replicates gives (4)4 + 3*4 = 268 runs • Echip achieves almost the same accuracy with 23 runs!

  36. Summary • Single variables use ANOVA to check models • Multivariable problems • Use design of experiments to see which variables are important (2n) designs • Software can simplify runs • Use variances to fit models (automatic in software)

  37. Class Question • What did you learn new today?

More Related