HRSG Modeling for Process Control - PowerPoint PPT Presentation

1 / 25
About This Presentation
Title:

HRSG Modeling for Process Control

Description:

HRSG Modeling for Process Control – PowerPoint PPT presentation

Number of Views:377
Avg rating:3.0/5.0
Slides: 26
Provided by: tommca
Category:

less

Transcript and Presenter's Notes

Title: HRSG Modeling for Process Control


1
HRSG Process Control Modeling Tom McAuliffe
Dec 3rd 2008 POWERGEN Gas Turbine
Technology HRSG
  • Valve Linearization
  • IP Bypass Feedforward Algorithm
  • HRSG Superheater Model

To hear sound please view presentation as a Slide
Show
2
Control Valve Linearization
  • Most PIDs Control Valves
  • PID Control Assumes Process Linearity
  • Cv DP are Non-Linear
  • Transition Between Valves in Parallel

3
Control Valve Linearization - Example
4
Small Valve Open Desired CV
Cv Curvefits
5
Then, Big Vv Open Desired Cv
6
Valve linearization In Practice
7
IP Bypass Feedforward Feedback
  • Plenty of Superheat lt 40 Flow
  • Poor Attemperation Flow Accuracy
  • Not Enough Superheat gt 40 Flow
  • Good Attemperation Flow Accuracy
  • Feedback to Feedforward Switch
  • _at_ 40 Sparger Pressure (40 Flow)

8
IP Bypass Problem Too Little Superheat
9
Feedforward Algorithm
10
IP Bypass Feed-forward Equation
  • Condenser Requires
  • hstm lt 2777 kJ/kg (1194 BTU/lbm)

Attemperator Feed-forward Spray Setpoint
Sparger Mass Flow Rate
11
Bypass Flow, Only Need One Data Point
Need Sparger Flow at a Pressure, _at_ Required
Enthalpy
Draw a Straight Line Through Zero
12
HRSG Superheater/Reheater Temperature
  • Critical to Efficiency/Reliability
  • Dearth of Model Info
  • Complex but Simple
  • Thermal Energy Storage (Heat Capacity)
  • Lumped Capacitance Does Not Apply
  • Biot Not lt 0.1 on Steam Side
  • Mass of Fins Mass of Tubes

13
HRSG Model Physical Data
14
Segments, Subsegments, Nodes
15
Example of Nodal Equation Development
Tube wall innermost nodes adjacent to outer
sub-segment boundaries experiences convection on
one surface and conduction on two surfaces.
Becomes Conceptually in Matrix Form
thermal diffusivity
16
Matrix Math is Key (Matlab)
  • 1000 seconds x 1000 ms/s 1,000,000 updates
    (for i 1 to)
  • About 8 hrs per test run (about 17 minutes of
    real time)
  • Could be worse
  • 60,722 nodal temperatures
  • 60,722 x 8 hrs 55 years
  • C matrices are diagonal
  • NSegments 120 for 60 foot duct, 137 for 67.5
    foot duct

17
HP Attemperation
18
RH Attemperation
19
HP Steam Flow Step Change
20
RH Steam Flow Step Change
21
CTG Exhaust Flow Step Change
22
CTG Exhaust Temperature Step Change
23
HRSG Model Observations
  • First Order Plus Deadtime (FODT)
  • Multi-Input Multi-Output (MIMO)
  • Inherently Stable (of course! )
  • MIMO IMC (Internal Model Control)
  • Gas Path Temperature Changes
  • Missing Steam Flow vs Heat Input

24
HRSG Model Future Additions
  • Incorporate Pressure Drop
  • Incorporate Evaporator Model Astrom Bell
    Drum Boiler Dynamics
  • MIMO IMC Manfred Morari Robust Process Control

25
Questions?
  • For Follow up Go To
  • www.powereng .com
  • Contact Hailey office
  • Ask for Tom (McAuliffe)
  • Or
  • Stop by our booth and ask for Tom
  • Booth 4332 Power Engineers
Write a Comment
User Comments (0)
About PowerShow.com