NeuCo/GE Boiler Optimization

Technology Overview and Results

July 18th, 2016

Imagination at work
 NeuCo, Inc./GE Power Collaboration
NeuCo - Power Optimization Market Leader
Only company 100% dedicated to power
optimization software
More than 120 active optimization systems
100% technology ownership and strong patent
position
Two U.S. Dept of Energy projects totaling $38M
investment

GE Power NeuCo / GE Power technology

Operating at Cholla and Jim Bridger
Leader in Industrial Internet design with defined the
technical roadmap for exploiting big data
Coal-fired boiler OEM and engineering services leader
Asset monitoring capability
2000+ units monitored remotely 24 X 7 X 365
Local Field Service presence

GE Proprietary Information
2
Boiler Optimization
Coal units are tightly regulated and unique
NeuCo and GE collaborating on advanced artificial intelligence and
controls to improve client profitability and reduce risk
We have proven products ready for implementation

BoilerOpt

CombustionOpt SootOpt

Optimizes fuel and air mixing Dynamically directs boiler cleaning

to reduce emissions and actions to achieve unit reliability,
improve efficiency efficiency and emissions goals

ProcessLink Platform GE Proprietary Information

3
 Comparison of Functionality for Historical and
Current Boiler Optimization Systems

Proprietary and Confidential GE Proprietary Information

4
 BoilerOpt Results

Proprietary and Confidential GE Proprietary Information

5
 CombustionOpt
Results achieved through Optimal mixing of fuel and air through MPC,
adaptive neural networks and condition-based rules
NOx reductions of 10-15%
Boiler efficiency increase of 0.5% - 1.5%
CO controlled to desired limit
Better ramping and load-following performance
Reduced opacity excursions
Better control of LOI
Better adherence to fan and mill amp limits
Improved situational awareness and process insight
Avoided tail-chasing behavior

GE Proprietary Information
Proprietary and Confidential GE Proprietary Information
Proprietary and Confidential GE Proprietary Information
Proprietary and Confidential GE Proprietary Information
 SootOpt
Real-time closed-loop optimization of boiler cleaning equipment using
expert rules, thermal calculations and neural networks

Reduced and more tightly controlled APH temperatures

Improved SH and RH steam temperature control
Reduced attemperation sprays
Heat rate reduction of 0.75% - 1.50%
Incremental NOx reduction of 2.5% - 5%
Avoided opacity excursions
Reduced blowing of 10% - 35%
Avoided thermal stress from blowing clean
surfaces
Fewer tube-leak failures
Improved situational awareness and process
GE Proprietary Information
insight
 Overview: How SootOpt Works

Zone Selection Blower Selection

Plant I/O Rules-based: Model-Prediction- and/or Rules-
Plant I/O
Fuzzy Based:
Heuristic Optimal Blower Start
First Principles Operating History
Operating History Effectiveness

Zones are defined according to

plant layout and sootblower
media piping One Blower Selection
Module per Zone

Proprietary and Confidential GE Proprietary Information

SootOpt Summary
Line

Blower
Selection
Table

Rule Selection Table

Proprietary and Confidential GE Proprietary Information

Proprietary and Confidential GE Proprietary Information
Strategic Investment with Measurable ROI
Reduces CO2 footprint
1% heat rate reduction saves fuel and reduces CO2 emissions 1%

CO2 reductions made after 2012 are recognized by EPA Clean Power Rule

Lowers cost and increases reliability of NOx / CO control

BART / NAAQS

Lower fleet NOx and required SCR / SNCR investment

EPA recognizes neural-net technology as a best practice technology

Required in many consent decrees

Included in DOE long term technical roadmap

Encouraged by MATS rule

MATS compliance risk and cost reduction

Reduce EPA data retention, required compliance tasks, frequency

GE Proprietary Information 14
14
 DTE Belle River Case Study

Imagination at work
 DTE Belle River Unit 2 Problem History

Imbalanced combustion creates areas of high slagging in SH and RH pendants

Multiple cleaning outages required every year

Boiler operated in high excess air (30-35%) comfort zone to minimize slagging
Less potential for low air / high slag areas
Lower FEGT
High NOx, Heat Rate, CO2 emissions, etc.
Very large imbalances still resulted in slagging
Limited success with boiler tuning
Little process data
Engineers / Operators stuck making
educated guesses
Boiler conditions constantly changing
Constant retuning required

Needed to find a way to exploit available

data and control functionality

GE Proprietary Information
 BR-2 Performance Test Results
Manual Tuning Neuco Tuning

Manual Neuco
Baseline Tuning Tuning Manual Manual Neuco Neuco
Heat Rate Heat Rate Heat Rate Tuning Tuning Tuning Tuning
Test Test Test Change Change Change Change
07/27/10 07/28/10 07/30/10 (Absolute) (Relative, %) (Absolute) (Relative, %)
Gross Load, MW 647.954 647.948 645.058 -0.006 0.00% -2.896 -0.45%
Net Load, MW 606.641 608.604 607.743 1.964 0.32% 1.102 0.18%
Auxiliary Power, MW 41.313 39.343 37.315 -1.970 -4.77% -3.998 -9.68%
Raw Net Unit Heat Rate (Heatloss),
BTU/kWhr 10517 10402 10331 -115 -1.10% -186.0 -1.77%
Corrected Net Unit Heat Rate (Heatloss),
BTU/kWhr 10393 10286 10224 -108 -1.0% -169.184 -1.63%
Net Unit Heat Rate (Input/Output),
BTU/kWhr 10493 10362 Not Avail. -131 -1.25% Not Avail. Not Avail.
Corrected Net Unit Heat Rate
(Input/Output), BTU/kWhr 10458 10358 Not Avail. -100 -0.96% Not Avail. Not Avail.
Nox, lb/MBTU 0.2513 0.2025 0.2010 -0.0488 -19.43% -0.050 -20.02%
CO, PPM 88 78 157 -10 -11.18% 68.200 77.18%
CO2 Intensity, Tons CO2/MWhr 1.069 1.047 1.043 -0.02 -2.06% -0.03 -2.43%
Total Boiler Air Flow, klb/hr 6313 5926 5483 -387 -6.13% -830 -13.14%
Average Excess O2, % 4.39% 3.23% 2.45% -1.15% -26.31% -0.019 -44.18%
Excess Air, % 30.50% 20.75% 15.12% 9.75% -31.97% -15.38% -50.43%

GE Proprietary Information
APS Four Corners Case Study

Imagination at work
 APS Four Corners Case Study
Initial success with CombustionOpt at Unit 3 led initially to deployment on
Units 1-2
All older front-fired 170-200 MW pulverized coal boilers

Success on Units 1-3 led to adding Units 4-5

Initial Objectives: HR & Opacity, NOx added later
Ten years of benefits led to widespread roll-out of new optimization
products
SootOpt, MOpt, POpt at all five Four Corners Units
COpt, SootOpt, POpt at all three APS-owned Cholla units

Also populating centralized fleet monitoring center

Extra eyes and subject experts for coal units
Remote monitoring and optimization for gas-fired units

Proprietary and Confidential GE Proprietary Information

 APS Four Corners Heat Rate
Improvement Over 7 Years

Focus on Focus on
CO/Opacity NOx

GE Proprietary Information
 Overall APS Fossil Objectives & Initiatives
Fossil-wide performance improvement team established.

Proprietary and Confidential GE Proprietary Information

 DTE Belle River Unit 2
B&W opposed wall-fired, balanced draft boiler
Built in 1984
Normal full load of 645 gross MW, Max load with overfire of 685 gross MW (turbine limited)
Pulverized coal from 8 B&W MPS-89 pulverizers, 7 operate during normal operation
5 burners per mill, 40 total
Originally 4 burner levels per wall, burners replaced with LNB and redistributed into 3 levels
Top level of burners replaced with OFA ports (1/3 and 2/3 control dampers in each port)
6 single-point extractive type O2 probes at economizer exit

LAND CO monitor and CEMS CO2 measurement in the stack

ABB INFI 90 distributed control system (DCS) with remote position control of
Burner secondary air shrouds and OFA dampers
OFA ratios / compartment dampers
Mill feeder speeds and primary air dampers
FD Fan inlet vanes for excess air/O2 control

PI system stores plant data

GE Proprietary Information
 CPS San Antonio Case Study

Imagination at work
 Optimization History at CPS Energy
CombustionOpt initially installed on all existing coal-fired
units
Spruce: 546 MW CE t-fired w/Ovation DCS
Deely: 2 x 446 MW CE t-fired w/Honeywell DCS

First installed at Spruce 1 in 2001

Installed at both Deely units in 2004
Primary objective of NOx reduction
Multivariate predictive control added in 2007
Additional objectives addressed through MPC
Explicit steam temperature control
Minimize attemperation sprays
Incremental heat rate and NOx reduction
SootOpt added to Spruce in 2010 and Deely in 2011
Spruce 2 project now getting underway GE Proprietary Information
J.K. Spruce Power Plant

Calaveras Power Plant Complex, South East of San

Antonio, TX
J.K. Spruce 1 commissioned in 1990
600 MW Alstom-CE T-fired unit
Emerson Ovation DCS
Spruce 2 commissioned 2011: 750 MW CE T-fired boiler
with SCR

Proprietary and Confidential GE Proprietary Information

25
Initial CombustionOpt NOx Results at JK Spruce

NOx Frequency Distribution- JK Spruce Unit 1

35.00%

30.00%

25.00%
Percentage Time

20.00%
May to Sept 2001
PL On - Jan-2002
15.00%

10.00%

5.00%

0.00%
0.09 0.11 0.13 0.15 0.17 0.19 0.21 0.23
NOx (lb/mmbtu)

Proprietary and Confidential GE Proprietary Information

26
BoilerOpt Impact on Spruce 1 NOx

Objective:
<0.125
Average = 0.132 Average = 0.122

GE Proprietary Information
Proprietary and
Spruce 1 East APH Temp SootOpt
Off vs. On

GE Proprietary Information
 BoilerOpt Impact on Spruce 1 Heat Rate

Objective:
Down
Average = 9856.6 Average = 9799.9

GE Proprietary Information
Proprietary and
Spruce 1 BoilerOpt
KPI Summary

Spruce Unit 1
KPI Units OFF ON Delta (ON - OFF) Delta (%) +
Gross MW MW 571.15 564.57
Heat Rate Btu/kWh 9661.32 9556.74 -104.58 -1.08%
Blr Eff % 83.99 84.77 0.78 0.93%
NOx #/MMBtu 0.127 0.121 -0.006 -4.72%
CO ppm 107.06 123.06 16 14.94%
O2 Avg % 2.277 1.912 -0.365 -16.03%
RH Temp E degF 1003.64 1003.88 0.24 0.02%
RH Temp W degF 1002.96 1003.33 0.37 0.04%
RH Spray klb/h 91.43 87.54 -3.89 -4.25%
SH Temp E degF 1003.26 1004.12 0.86 0.09%
SH Temp W degF 1000.16 1003.84 3.68 0.37%
SH Spray klb/h 14.9 21.89 6.99 46.91%

Proprietary and Confidential GE Proprietary Information

30
 JT Deely Power Plant
Calaveras Power Plant Complex, South East of San
Antonio, TX.
J.T. Deely 1 & 2 plants commissioned in 1977 & 1978
Boilers are identical 446 MW Alstom-CE T-fired units
Common stack
Low-NOx burner tips and SOFA
Honeywell DCS

GE Proprietary Information
 Adding MPC to CombustionOpt at Deely
OFA Bias
Fan Bias

Neural Aux Air Bias

Network OFA Bias
Optimizer
Feeder Bias
CEM NOx
F/A Damper Bias
Boiler Outlet CO
Mill Air Bias

Zolo CO paths
Zolo O2 paths Profile

Zolo Temp paths

Reheat Temp Model O2 Bias

and Sprays Predictive
Control Burner Tilts
Superheat Temp
and Sprays SOFA Tilts

GE Proprietary Information
On vs. Off Results after Adding MPC at Deely

GE Proprietary Information
Deely 2 SCR Inlet Temps
SootOpt Off vs. On

GE Proprietary Information
 SCR/SNCR Systems & Optimization
Reduce Reagent Usage
Lengthen Maintenance Intervals
Avoid Ammonia Slip
Reduce risk of Ammonium Bisulfate & Sulfur Trioxide deposits
Control Blue-Plume Opacity Excursions
Tighter, condition-based gas temperature control
Better Manage Interactions

GE Proprietary Information
Impact on NH3 Usage

WITHOUT Optimization, 400-600 klb/hr WITH Optimization, 300-400

GE Proprietary klb/hr
Information
NH3 flow needed to meet NOx target NH3 flow needed to meet NOx target
 CombustionOpt Ammonia Reduction

Ozone Ammonia Flow - 2005 Ozone Ammonia Flow - 2006

800 0.160 800 0.160

700 0.140 700 0.140

Gross Mw / NH3 Flow (#/Hr)

Gross Mw / NH3 Flow (#/Hr)

600 0.120 600 0.120

NOx Rate (#/MMBtu)

NOx Rate (#/MMBtu)

500 0.100 500 0.100

400 0.080 400 0.080

300 0.060 300 0.060

200 0.040 200 0.040

100 0.020 100 0.020

0 0.000 0 0.000
11 05

12 05

5
27 05
11 05
23 05
7/ 005

20 05
8/ 005

15 05
26 05
13 05
25 05

06 06 006 006 006 006 006 006 006 006 006 006
21 0

00

20 20
5/ /2 0

7/ 2 0

8/ 2 0
5/ / 20
6/ / 20
6/ / 20

7/ / 20

9/ / 20
9/ / 20
9/ / 20
9/ / 20

2 2 2 2 2 2 2
2/ 22/ /1/2 20/ 14/ 22/ /4/2 12/ 23/ /2/2 14/ 28/
/2

/2

/2
1/

5/
1

1
5/ 5/ 6 6/ 7/ 7/ 8 8/ 8/ 9 9/ 9/
5/

BA1 Gross Load BA1 NH3 Flow 1A BA1 NH3 Flow 1B BA1 CEM S NOx BA1 Gross Load BA1 NH3 Flow 1A BA1 NH3 Flow 1B BA1 CEM S NOx

GE Proprietary Information
 Boiler and Back-end Processes
Process Causality Chain
Exogenous & Manipulated Process Process Cost
Disturbance Variables States Performance Streams
Variables

NH3 Injection CO

Fuel-Air Mix Fuel Usage

NOx
Fuel Quality Fuel Flow

Heat Transfer SO2 Reagent Usage

Ambient Temp Fuel Distribution

Gross HR
SH Temp Catalyst Life

Barometric Air Flow Ash Deposition

Pressure
SH Sprays FGD Sorbent
Usage
Carbon-in-Ash
Inlet H20 Temp Air Distribution
RH Temp DSI Usage
Slagging
MW Demand Boiler Cleaning
RH Sprays ACI Usage
Erosion
Ramp Rate Temp Controls
Gas Temps Net Ash Disposal
Sorbent Injection Corrosion
Forced Outage
Thermal Stress Rate

Proprietary and Confidential GE Proprietary Information

38
 Xcel Energy Case Study

Imagination at work
 Overall Xcel Optimization Initiative

Xcel SPS decided to adopt boiler optimization software (fuel-air and heat
transfer components)
Tested NeuCos BoilerOpt system at Tolk station early-mid 2009
Applied to Harrington Unit 3 in October 2009
Rolled out to Harrington Unit 2 and Unit 1
Xcel subsequently standardized on BoilerOpt as part of their fleet-wide
boiler reliability program, with additional installations in Minnesota and
Colorado completed and underway now

GE Proprietary Information
 NeuCo at Xcel Energy
Project Timeline
Timeline (Approximate) of NeuCo Installations at Xcel

2008 2010 2012 2014

CombustionOpt Project Customer Support Services

Tolk 1 SootOpt

Tolk 2

Harrington 1

T-Fired
Harrington 2

Harrington 3
BoilerOpt
Sherco 1 (typ.)

Sherco 2

Pawnee Wall-Fired

GE Proprietary Information

Proprietary and Confidential 41

BoilerOpt at Xcel Harrington

Units 1, 2 and 3 similar units:

CE T-Fired boilers
360 MW each
Often on AGC between 180 and 360
5 Pulverizers
LNBs, CCOFA and SOFA
Foxboro I/A DCS

Baghouses on U2 & U3
ESP on U1

Proprietary and Confidential GE Proprietary Information

42
 Harrington Unit 2 NOx ON/OFF

0.16 0.13 14%

Reduction

Proprietary and Confidential GE Proprietary Information

43
 Harrington Unit 3
Long-Term Trends

No wind, Wind/SPP dispatch, Wind/SPP dispatch,

no Optimization no Optimization with Optimization

Proprietary and Confidential GE Proprietary Information

44
Continuous Improvement
Xcel Tolk example

Proprietary and Confidential GE Proprietary Information

45
Using new knowledge

After reviewing the results and our recommendations with

plant
Adjustments were made to the existing Zone/blower map in SootOpt to
account for blowers affecting both RH and Platen sections.
Adjustment were made to deal with a cross-over between DivSH and
PlatenSH that had a noticeable effect
Adjusted existing rules and thresholds
Added rules that inhibit cleaning when section deltaTemp is at or above
its normal mean, given operating conditions
Added rules that propose cleaning when section deltaTemp is
sufficiently below its normal mean

Proprietary and Confidential GE Proprietary Information

46
Recent Results Tolk Unit 2:
Blower Counts

- 20%

Proprietary and Confidential GE Proprietary Information

47
Recent Results: NOx, CO

GE Proprietary Information
Recent Results: RH Temps

GE Proprietary Information
Recent Results: Heat Loss Index

GE Proprietary Information
 Standardize to Improve Fleet Performance
Best operator performance 24 X 7

Fleetwide KPIs and best practices instantiated in

software
KPI tracking common to all units

Helps newer operators to master operating processes

Lower NOx and heat rate with less stress improves

reliability
Reduces tube leak related unplanned outages

Minimizes volatility of operations Chehalis Power Project and Currant

Creek Power Plant operation supported
O&M flexibility and lower O&M cost by GE monitoring center ioutside
Atlanta, Ga.
Lengthens interval between required MATS testing, tuning
and repair
Advanced analytics enable early problem detection
Secure web interface enables leverage on GE engineering
expertise

GE Proprietary Information 51
51
Favorable Regulatory Treatment
Lowers cost to end customer
Reduces fuel and reagent costs
More discretion in O&M expense and frequency
Reduces CO2 and long-term Clean Power Rule compliance cost

In line with EPA approved technologies

In rules, consent decrees and long term technology roadmap

Offered with flexible commercial terms to meet PUC requirements

Capital purchase of perpetual license
Annual service fee
Lagged benefits sharing enables payments after benefits accrued
NERC CIP compliant installation / operation

GE Proprietary Information 52
52