CHEE 209 Analysis of Process Data

1
CHEE 209Analysis of Process Data

• Tom Harris
• Fall 2009

The original compilation of these notes by Prof.
Kim McAuley
2
Role of Statistical Methods

• Summarizing data
• using a few key numbers or graphs
• uncover or reveal complex patterns
• Decision-making in the presence of uncertainty
• Has the concentration of a contaminant in an
  effluent stream changed appreciably?
• Has the quality of our polyurethane foam
  improved?
• Is the quality of a mineral deposit sufficient
  to justify investments?
• Use confidence limits and hypothesis tests and
  control charts to find answers to these
  questions.

3
Role of Statistical Methods

• 3. Identifying relationships in data
• Designing experiments to uncover relationships
  among variables,
• Fitting empirical models between reactor
  temperature and product yield
• Development of calibration methods for chemical
  property analysis
• Test suitability of various models

4
Why do need statistical methods?

• Presence of Process or Inherent Variability
• Price of Platinum
• Temperature Measurements
• Moisture Content of a Paper Product

5
Sources of Process Variability

• instrumentation and measurements
• operator interventions
• disturbances
• from raw materials or upstream units
• Temperature or composition of feed stream
• ambient conditions
• Air temperature
• Cooling water temperature
• external factors demand, speculation

6
Types of Data

• Discrete
• number of defective items, number of people
• colours
• taste
• Continuous
• temperature
• pressure
• composition

Combination discrete continuous
7
Types of Data

• May be collected with respect to time, i.e.
  composition measurements every hour, minute or
  second
• Spatial component
• Collected with respect to distance
• 2 or 3 D

8
Data Collection

• Passive Collection
• record results without actively intervening in
  the operation of the process
• historical databases of operational or
  geophysical data
• Active Collection Intervention
• do a planned experiment
• guarantees a cause and effect relationship
• Example police officers and crime

9
Value of statistical methods

• Established body of both theory practice that
  provides a scientific systematic approach for
• collection of data
• methods for analysis and interpretation
• Used to assist in decision making (statistics
  doesnt replace sound engineering judgment and
  knowledge)
• Universal language that transcends specific
  disciplines, i.e. engineering, business, law,
  medicine

10
Course Overview

• describing data
• describing variability
• probability
• decision-making in the presence of uncertainty
• hypothesis tests and confidence intervals
• statistical quality control
• empirical modeling
• linear regression

Examples from chemical engineering, chemistry
geosciences
11
Course Objectives

• Understand motivation for use of statistics
• Proficient and confident in using a number of
  introductory techniques using hand and
  computer-based calculations
• Develop a critical awareness regarding
  statistical thinking which is important for
  your responsibilities as engineers and citizens.