E3 Boiler Combustion Optimization System

E3

Applying
ApplyingArtificial
ArtificialIntelligence,
Intelligence,Advanced
AdvancedControl,
Control,and
andOptimization
Optimization
Technologies
TechnologiesIn InCoal
CoalFired
FiredGenerating
GeneratingUnits
Units

Jianhu
JianhuJia,
Jia,Ph.
Ph.DD

Pegasus
PegasusTechnologies,
Technologies,Inc
Inc

Confidential and Proprietary Information

Leaders in Combustion Optimization for the Power Industry
 Pegasus Technologies
Headquarters in Ohio, USA

Globally established
North America
Asia
Europe

Experienced Staff for Power Market

Engineers: Combustion, Advanced Controls, IT

Confidential and Proprietary Information

 Power Generating Unit

Complex System, Consists of many sub-

system ,
Air/Fuel/Gas /
Steam Cycle
BOP
Hundreds of pieces of Equipment make up
these systems

Confidential and Proprietary Information

 Current DCS System
DCS

Efficiently Control the Unit

Doing Good in Managing Safety Interlocks
and Operating Sequence

Based mainly on Feedback Controls PID

Inherently Single Input/Single Output

Confidential and Proprietary Information

Keeping Unit at Optimal Condition
Is a Complex Task

Coupling Among Various Parameters

Unmeasured Disturbances
Transients
Nonlinearities

Confidential and Proprietary Information

 Combustion Dynamics
The Complex Relationships of Multiple Variables That Are Continuously Changing
Dramatically Impact Combustion

Superheat and Reheat Temp. Different Operators

Furnace DP Instruments Failure / Drift
/

Equipment Conditions
SA Temp. and Flow

Weather
Flue Gas NOx
Humidity
PA Temp. and Flow Quality/Type of Coal
/

Excess O2 Atmospheric Conditions

Burner Tilts (T-Fired)

()
Confidential and Proprietary Information
 Multiple Objectives
Combustion is Continuously Optimized Based On Changing
Plant Conditions and your Unit Specific Operation Objectives...

NOx
Heat Rate
Capacity
Opacity
LOI
Steam Temperatures

?
Objectives can be weighted for different seasonal operations.
operations

Confidential and Proprietary Information

Zone 1 Zone 2 Zone 3

Boiler Efficiency

Combustible Losses
Gas Losses

Low High

Excess Air
Confidential and Proprietary Information
 Power Plant Process Example

The following are NOx and CO responses with respect to each of the SOFA Dampers at
a specific Windbox/Furnace DP. /

Confidential and Proprietary Information

Pegasus Combustion Optimization
Technology
Pegasus

Specially Designed to Optimize

Power Plants

Confidential and Proprietary Information

 Intelligent systems can be utilized to control the plant for

Reduced Coal Consumption

Reduced Environmental Emissions (NOx, SO2, CO2)
(NOx, SO2, CO2)
Ash (C content, heat absorption, slag, fouling) (
)
Fuel (quality, blending) ( )
Reduction of Equipment Failures / Unplanned Outages

Sootblower (sequencing, optimization) ( )

Optimum Benefit through Closed-Loop Operation

Confidential and Proprietary Information

E3 Boiler Optimization
E3

Control Commands to DCS

Unit Information

Confidential and Proprietary Information

 Neural Networks
Main Features of Neural Net :

Very Powerful Universal Approximating Modeling Tool

Strong Abilities to Capture Non-Linear Relationships

F to Model Complex (multi-dimensional) relationships
1 ()

Confidential and Proprietary Information

Artificial Neural Networks

Confidential and Proprietary Information

 Model Predictive Control
Use first step of horizon
MV prediction

MV measured value

MV set point

CV1 measured value Perfecter optimizes future trajectories

using an internal models
Perfecter

CV1 prediction 1
current time indicator

CV1 desired value 1

CV2 prediction 2

CV2 Upper Fuzzy Limit

2

Confidential and Proprietary Information

 Combustion Optimization Challenge

To achieve the optimal combustion model - the system

manages and analyzes up to 1500 or more inputs
-
The system accurately models the complex
interrelationships among 5-60 controllable parameters

Optimization occurs within constraints of the operating
environment
The system responds to changing environmental and
operational conditions

Confidential and Proprietary Information

 Pegasus Solution for Dynamic Optimization
Pegasus

Pegasuss Proven Dynamic Systems

Pegasus
o Model are developed for optimization at various loads and conditions

o Model Predictive Control technology to control the unit at optimal condition
on a continuous bases
o System will learn and adapt to new conditions and find new solutions for
optimized results
o Global optimization for improved control of operating conditions

Note: Retraining still required for Major control or boiler modifications

: ,

Confidential and Proprietary Information

 Pegasus Closed-Loop
Up to 1500 Current Inputs From DCS
Continuously Collected and Analyzed
,

5-60 Manipulated Output Settings Sent

Pegasus
To DCS on Fixed Interval
Computer ,,DCS

Control Room
DCS
Operator Field Sensors and Drives

Confidential and Proprietary Information

 Systems Integration
DCS
Feedforward
or
Error
Signal
Operator Bias
or
DCS Fixed Bias
Operator Setpoint/Demand
or
DCS Fixed Setpoint/Demand

f(x) Power Perfecter

or Bias
PID SWITCH

Power Perfecter
Setpoint/Demand
SWITCH

SUMMER
PID Setpoint
(or Demand)

Bias
+
PID Setpoint
(or Demand)

Confidential and Proprietary Information

Example1- Xibaipo Unit #3
- 3

Confidential and Proprietary Information

 Unit Information
300MW 30
Front and Rear Wall Fired
Indirectly Fired
Compartmentrized Windbox
WDPF DCS DCS

Confidential and Proprietary Information

Boiler Configuration

Confidential and Proprietary Information

Mill Configuration

Confidential and Proprietary Information

Manipulated Variables
1 FD Fan Bias
2 2nd Air Pressure Bias
3 FU 2nd Air Damper Bias
4 FM 2nd Air Damper Bias
5 FD 2nd Air Damper Bias
6 BU 2nd Air Damper Bias
7 BM 2nd Air Damper Bias
8 BD 2nd Air Damper Bias
9 PA Pressure Bias
10 PA Fan Bias
11 Pri SH Temp SP A Bias
12 Pri SH Temp SP B Bias
13 2nd Stage SHT SP A Bias
14 2nd Stage SHT SP B Bias
15 RH Temp SP A Bias
16 RH Temp SP B Bias

Confidential and Proprietary Information

Controlled Variables
1 Nox 13 PA_AMP_DIF
2 O2_A 14 Pr_Spry_A
3 O2_B 15 Pr_Spry_B
4 Total_SA_Air 16 SH_Spry_A
5 FU_FAR 17 SH_Spry_B
6 FM_FAR 18 RH_Spry_A
7 FD_FAR 19 RH_Spry_B
8 BU_FAR 20 FD__Dmd
9 BM_FAR 21 ID__Dmd
10 BD_FAR 22 FD__Amps
11 O2_DIF 23 ID__Amps
12 FD_AMP_DIF 24 PerfIndex

Confidential and Proprietary Information

Pegasus Interface Design
DCS

Confidential and Proprietary Information

Data Link

Confidential and Proprietary Information

Optimal Controller - Data Flow Diagram
-

Confidential and Proprietary Information

Pegasus Optimal Controller - GUI

Confidential and Proprietary Information

Boiler Combustion Optimization System - GUI

Confidential and Proprietary Information

 NOx Comparison

NOx Comparison
- Xibaipo #3 Before vs After Optimization
700
NOx (ppm, normalized to 7% O2)

600

500

400

Averaged NOx Before Optimization

300 Avergaed NOx after Optimization

200
200 220 240 260 280 300 320
Load (MW)

Confidential and Proprietary Information

 Perfindex Comparison

PerfIndex Comparison
- Xibaipo #3 Before vs After Optimization
1000

900

800

700
Perfindex

600

500 Averaged Perfindex Before Optimization

Avergae Perfindex after Optimization
400

300
200 220 240 260 280 300 320
Load (MW)

Confidential and Proprietary Information

 Improvements

NOx improvement 10.5%

Heatrate improvement 0.63% to 0.95%

Perfindex improvement 0.79%

Confidential and Proprietary Information

Questions?

Confidential and Proprietary Information