Teaching Improved Methods of Tuning and Adjusting HVAC Control Systems

1
Teaching Improved Methods of Tuning and Adjusting
HVAC Control Systems
A National Science Foundation Grant Grant No.
NSF DUE 0202131 Presented at the 2006 ASEE Annual
Conference Exposition Chicago, IL

• Professor Russell Marcks, PE Principal
  Investigator and Presenter
• Assistant Professor Larraine Kapka, PE
  HVAC Grant Project Manager
• Dr. Alan Watton, PE Principal
  Investigator

2
Project History

• Early 1990s NSF-funded project to build lab
  demos to study control loop dynamic response
• In a national seminar, we demonstrated a
  graphical evaluation method using a flat-bed
  plotter to collect loop response data
• Seminar attendees determined this quite useful if
  instrumentation was portable

3
Project Goals

• Robust method of field data collection
• Cost less than 1000
• Able to be used by trained technician rather than
  control engineer
• Instructional package to educate and train HVAC
  students in control loop tuning

4
Background

• HVAC industry has frequently approached the
  problem of tuning control loops informally
• Excerpt from the manual of a well-known HVAC
  control manufacturer
• Adjust reset time (Tr), according to the job
  drawings, with reset time adjustment knob.
  Decrease setting until system becomes unstable.
  Increase setting slightly until system becomes
  stable.

5
Background

• Process control industry regularly engages in
  evaluation of control loop response
• Highly sophisticated tuning software
• Common communications protocols allow networking
  of different controllers

6
Why is HVAC behind?

• Most HVAC control technicians and engineers do
  not understand formal loop tuning
• No common communications protocol exists to
  easily apply software monitoring
• Owners resistant to added costs for software
  monitoring as is available in process industry

7
Why is HVAC behind?

• Attitudes
• Its just temperature close enough!
• Who cares if it doesnt react quickly?
• Attitudes are changing
• Energy costs
• ASHRAE ventilation requirements
• A move to formalized building commissioning
• LEED and Green

8
The Current State

• According to some sources
• 30 of controllers are operated in manual
• 65 of controllers are poorly tuned or de-tuned
  to mask control related problems
• 50 of actuators, transducers and positioners are
  improperly calibrated
• 20 of control systems are not properly
  configured to meet control objectives
• Source Tecmation, Inc.

9
Why Tune?

• Properly tuned system reacts when process
  requires and not over-react to process noise
• Formalized tuning process provides visual
  evidence of a variety of control problems
• Reduction of OM costs

10
But We Have Autotune!

• If it works, use it
• Many controllers do not autotune
• Requires significant processor power
• Autotune assumes system is linear HVAC control
  systems are not
• Proposed method provides a visual confirmation

11
Some OM Costs

• Energy Use
• Poorly tuned loops wear valves
• Valves represent up to 75 of system first cost,
  but 90 of loop maintenance cost
• Valve moves 15 to 200 times/day in well tuned
  loop, 50 times as much in poorly tuned loop
• Air typically costs 100/yr/scfm
• Source Top Control, Inc., 2002

12
Potential Valve Savings
Based on 0.06/kWh Source Top Control, Inc.,
2002
13
DDC Controls in Oregon Public Schools

• Manual operation and bypass of 144,000 DDC
  system
• 77,500 DDC system operates 24 hrs/day,
  over-pressurized building, poor room control
• 100,000 in energy efficiency retrofits
  (including DDC system), utility bill increased
  from 11,000 to 17,000/yr

Source Oregon Office of Energy from report on
DDC systems in Oregon public schools
14
DDC Controls in Oregon Public Schools

• Estimate one-sixth of Oregon K-12 schools
  have dysfunctional DDC systemsBy fixing DDC
  design, installation and user errorscut their
  energy costs by a total of more than 1 million
  per yearincrease building comfort and reduce the
  amount of staff time on jerry-rigging the DDC and
  HVAC systems to work correctly.

Source Oregon Office of Energy from report on
DDC systems in Oregon public schools
15
A New Approach

• Simple, inexpensive method of collecting data in
  the field
• Analysis results in a set of tuning parameters
  allowing for optimized control loop response or
• if analysis shows difficulties then take back to
  design office for further study and resolution
• Overall goal is to get large-scale control action
  to follow the desired operating pattern

16
Hardware Requirements

• Calculator programmable and flexible
• Memory to collect and store data
• Variable sampling rate
• Accepts two synchronized inputs
• Analyze data directly or
• easily download data to computer
• Batteries or AC power

17
Simplified Data Collection

• Handheld and portable
• Easy to program / setup / use
• Inexpensive
• Provides data plot
• Controlled data collection

18
Test System
19
Data for Loop Tuning
20
Data Analysis

• May use any desired tuning method
• Ziegler-Nichols (Open Closed Loop)
• Cohen-Coon (Open Loop)
• Lambda (Dead Time Dominant)
• Current analysis is a graphical hand analysis
• Easily adaptable to Excel or MathCAD for further
  automation of analysis procedure

21
TI-83/84 and LabPro

• Pros
• Easy to use
• Easy to program
• Inexpensive
• Cons
• Should learn to program
• Limited active memory
• Educational use only

22
Casio CFX-9850GC and EA-200

• Pros
• Color Graphing
• More active RAM than TI-83
• Easy to use and program
• Cons
• More expensive than Vernier
• No storage on Calculator
• Slow Processor
• Not as well supported in the US as TI-83

23
Casio FX-9860G SD and EA-200

• Pros
• More active RAM than TI-83
• 1.5 MB total on-board RAM
• Accepts up to 1 GB SD Card
• Built-in Spreadsheet
• Easy to use and program
• Cons
• More expensive than Vernier
• Not as well supported in the US as TI-83

24
Fourier Systems

• Pros
• Easy to use - Intuitive
• No programming
• Ready out of the box
• Industrial use
• Full range of process signals
• Cons
• Somewhat more expensive
• No onboard calculation

• MultiLogPro DAQPRO

25
Fourier Systems

• Pros
• Familiar Windows Interface
• Tablet PC with datalogger
• Spreadsheet Calculations
• Can connect Keyboard, et.al.
• Cons
• Targeted to K-12
• VERY new

• Nova 5000EX

26
HP and Firmware Systems

• Pros
• More capable than the TI-83
• Easy to use/Program
• Designed for industry first
• Cons
• Must learn to program
• Limited sensor availability
• Availability

27
Field Application Issues

• Accessibility of control signals
• Manufacturer warranty
• Lack of adequate time
• Elimination of non-linearities and anomalies
• Connection to pneumatic systems

28
Education Component
29
Education Component

• Weve successfully used these procedures to teach
  loop tuning in the classroom with 2-year HVAC
  students

30
Education Component

• Learning materials to provide a full course in
  introductory control including loop tuning are
  nearly complete

• By the end of June, 2006, complete learning
  materials will be available, on CD-ROM, through
  the National Center for Manufacturing Education
  (NCME)

31
Conclusion

• Inexpensive data collection is available
• Based on raw data collection so independent of
  communications protocols
• Tuning methods are easily learned
• Method provides documentation
• Can aid in identifying problems otherwise
  difficult to uncover
• Can reduce energy and OM costs

32
For Educational Materials

• www.hvacrcool.orgOn-line soon

33
For Further Information

• See us Tuesday at Booth 825, NSF Resource Centers
• Browse To
• www.ncmeresource.org
• www.hvacrcool.org
• Contact
• russell.marcks_at_sinclair.edu
• larraine.kapka_at_sinclair.edu