BR 1844 Babcock Mercury

8/3/2019 BR 1844 Babcock Mercury

http://slidepdf.com/reader/full/br-1844-babcock-mercury 1/10

Multi-Technology, Mercury Abatement Program, Implementation and Performanceat Nova Scotia Power

Technical PaperBR-1844

Authors:

D. McLellan

Nova Scoa Power Incorporated

Haifax, Nova Scoa, Canada

B.J. Jankura

Babcock & Wilcox

Power Generaon Group, Inc.

Barberton, Ohio, U.S.A.

M. MacKenzie

R. Ojanpera

Babcock & Wilcox Canda

Cambridge, Ontario, Canada

Presented to:EPRI Power Plant Air Pollutant

Control Mega Symposium

Date:

August 30 - Sept. 2, 2010

Locaon:

Balmore, Maryland, U.S.A.



Babcock & Wilcox Power Generation Group 1

Multi-Technology, Mercury Abatement Program, Implementation and Performance at Nova Scotia Power

B.J. Jankura

Babcock & Wilcox

Power Generation Group, Inc.

Barberton, Ohio, U.S.A.

BR-1844

Abstract

Nova Scotia Power Incorporated (NSPI) has installed

mercury capture technologies on seven coal-red boilers to

comply with provincial regulations for mercury emissions.

To achieve with the 2010 provincial cap of 65 kg/yr, a

multi-technology approach consisting of enhanced mercury

oxidation followed by halogenated activated carbon injec-

tion, has targeted 70% coal-to-stack reduction from existing

total mercury emissions. This paper discusses programimplementation and performance test results at seven NSPI

boilers using the combination of calcium chloride injection

technology with Powdered Activated Carbon (PAC) injec-

tion upstream of the particulate collection devices. Both

systems were installed on several different plant congura-

tions, including both front and tangentially red boilers, with

electrostatic precipitators and one unit with a baghouse. The

design and performance issues with novel “hot-side” PAC

injection lances before the air preheater and performance

with Colombian, bituminous coal are included.

Introduction Nova Scotia Power Incorporated is a regulated, vertically-

integrated, electric utility. The company supplies over 97%

of the generation, transmission and distribution of electrical

power to 460,000 customers in the province of Nova Scotia,

Canada. NSPI owns 2,293 megawatts (MW) of generation

capacity, fuelled by a mix of coal and petroleum coke (55%),

oil and natural gas (27%) and hydro, tidal and wind (18%).

Mercury regulation development

The development of a mercury regulation in Nova Scotia,

leading to the implementation of mercury emission controls

at Nova Scotia Power, was over a relatively short timeframe.

Canadian coal red utilities participated in a collective

program administered through the Canadian Electricity As-

sociation to gain an understanding of mercury measurement

and capture (REF1). The information gathered in this program

was then reviewed and supported decision making in theCanada Wide Standard process completed by the Canadian

Council of Ministers of the Environment (REF2).

In 2005, a 168 kg annual mercury cap came into effect

in Nova Scotia. This was achieved through fuel switching

and intensive monitoring. However, following the release

of the Canada Wide Standard for Mercury and subsequent

revision to the Nova Scotia Air Quality Regulations reducing

the annual cap to 65 kg, the required reduction could not

be met by fuel switching alone (REF3). At this point, NSPI

completed a test program to assess if the 2010 cap could be

met by commercially available control technology.

Following the testing, a procurement process and work

order submission to the regulator were immediately initiatedto meet the aggressive implementation schedule. Figure 1

gives an overview of the events leading to the implementa-

tion of mercury control systems.

Summary of NSPI units with installedsystems

There are four solid-fuel burning thermal stations in

the NSPI eet, and it was at three of these stations that the

mercury control equipment was installed. This includes a

Presented to:

EPRI Mega Symposium 2010

August 30 - Sept. 2, 2010Baltimore, Maryland, U.S.A.

M. MacKenzie

R. Ojanpera

Babcock & Wilcox Canada

Cambridge, Ontario, Canada

D. McLellan

Nova Scotia Power Incorporated

Halifax, Nova Scotia, Canada



2 Babcock & Wilcox Power Generation Group

total of seven boilers at the three stations. Relevant details

of the plants are discussed below and summarized in Table 1.

Lingan - This plant is the largest in the thermal eet

and currently operates four (4) 150 MW tangentially red

units. The units are of the same manufacture and vintage,

and have been retrotted with low NOx systems, including

separated overre air. Particulate collection is by means of

an electrostatic precipitator.

Point Tupper - Pt Tupper Unit 2, the remaining unitoperating at the plant, was converted from a heavy oil red

unit in the 1980’s. Unit 2 is a 150 MW tangentially-red

boiler and has also been retrot with a low NO x combustion

system including separated overre air. Particulate collec-

tion is by means of an electrostatic precipitator.

Trenton - This plant currently operates two (2) units.

Unit 5 is 150 MW, constructed in the late 1960’s, and has

the original circular register burners. A baghouse was added

in 2009 to supplement the original electrostatic precipita-

tor. Unit 6, with nameplate capacity of 160MW, went into

service in 1991 and has an electrostatic precipitator for

particulate collection.

Test fuel

Nova Scotia Power has burned a variety of fuels and

fuel blends over the last number of years as local supplies

became less available. These blends are tested and adjusted

accordingly to meet environmental and operational con-

straints. When identifying the fuel that would be used to

complete testing on the new mercury abatement systems, a

low-sulfur Colombian coal with a higher heating value av-

eraging around 10,300 Btu/lb was selected since it would be

readily available for the testing and used for the rst several

months of operation.

Mercury control technology selectionFollowing the revision of provincial regulations to set the

2010 mercury cap, NSPI completed a review of commer-

cially available technologies that could be used to meet the

required emission reduction. This review led to pilot scaletesting of activated carbon injection and sorbent enhance-

ment additives.

After a review of the test results, NSPI made the deci-

sion to initiate a procurement process that would lead to

the purchase and installation of seven mercury abatement

systems. Results indicated that activated carbon injection

could meet an emissions reduction of almost 70% and the

use of brominated activated carbon or sorbent enhancement

additives would provide additional margin to meet the

regulation. Also a consideration when selecting the control

Fig. 1 Timeline of development and regulatory milestones.

Fig. 2 MercPlus™ system storage system.




technology, were the long term opportunities to minimize

operation and maintenance costs and have exibility in the

use of future fuels at each station.

MercPlus™ system

Mercury control at the NSPI plants is accomplished by

a combination of two technologies. The rst technology is

the MercPlus™ front-end fuel additive system from Babcock

& Wilcox Power Generation Group, Inc. (B&W PGG) to

enhance the oxidation of mercury in the boiler. The system

is composed of liquid solution storage tanks, pumping/

metering skid(s), coal injectors, and DCS interface control-

lers. The storage tanks at NSPI are berglass construction

to allow for the various sizes and congurations required

at the multiple locations. There is one pumping/metering

skid for each boiler. The skids have multiple pumps with asingle pump dedicated to each coal feeder injection point.

The pumps are sized sufciently that should one pump need

to be taken down for service, the remaining pumps can au-

tomatically increase their rates to keep the overall injection

rate for each boiler at the proper level.

The MercPlus system sprays a CaCl2 solution on the coal

prior to combustion. This is advantageous because oxidized

mercury can be captured with relatively high efciency. The

additive is injected directly onto the coal as a 30% solution

of CaCl2 (20% chlorine), which is commercially readily

available from multiple suppliers. It is stored in tanks that

can be either inside the boiler building or outside without

freeze protection. Figure 2 shows pictures of the MercPlus

storage system. The physical properties of the solution of –55F freeze point do not require heat tracing on any of the

equipment or piping even in the very low temperatures typi-

cally experienced in the plant locations.

The solution is pumped by the skids from the storage

tanks to the coal feeder deck in the boiler building where it

is sprayed onto the coal in each of the operating coal feed-

ers upstream of the pulverizers. Figure 3 shows the solution

ow control skids.

Figure 4 shows the coal feeder injector piping. The in-

jection rate is metered to the feeders as a function of feeder

speed, which is proportional to coal ow, to achieve the de-

sired additive stoichiometry. As opposed to furnace reagent

injection, or coal silo injection, this is a very simple system,which affords fewer injection points, good reagent mixing,

even distribution among burners, and a quick response time.

The injection rate of the chloride solution additive was

automatically controlled by the DCS system with logic pro-

vided by B&W PGG. The equipment operator could directly

select the required injection rate for unattended operation.

The chloride addition rate is based on a ratio to the coal ow

measured in pounds of chlorine per million pounds of coal

to the boiler – dry basis (ppmd).

Table 1

Summary of Nova Scotia Plant Designs

Plant Lingan Point Tupper Trenton 5 Trenton 6

Unit Nameplate Capacity 4 @ 150 MW 1 @ 150 MW 1 @ 150 MW 1 @ 160 MW

Combustion System

Corner red,

low NOx with

Overre Air

Corner red, low NOx

with Overre Air

Front Wall red,

original burners

Front Wall red,

low NOx burners

and overre air

Particulate Collection Device(s) Cold Side ESP Cold Side ESPCold Side ESP

and BaghouseCold Side ESP

ESP Details

Total Collecting Area (sq. ft) 130,159 120,960 103,500 173,430

# of electrical elds 6 3 4 3

# of mechanical field 3 3 3 3

Baghouse Details

Air/Cloth Ratio -------- -------- 3.46 --------

Net Cloth Area (sq. ft.) -------- -------- 166,445 --------




Activated carbon addition system

Injection of Powdered Activated Carbon (PAC) into the

ue gas is the primary technology used for mercury control

at NSPI. Each boiler has a dedicated system which includes

storage silo, blower, feeder, piping, splitters, distribu-

tion bottles and lances. Each silo contains one redundant

blower (i.e. two blowers/silo, one operating at any point in

time) and eductor. Figure 5 shows the four PAC silos at the

Lingan Plant.

While the nominal ratings of the seven plants are similar,

there are four distinct plant designs which required speciclayout and equipment specications. A summary of the

signicant design parameters of the carbon injection equip-

ment is shown in Table 2.

PAC is pneumatically transferred from the silo to the

injectors through a series of rigid piping and exible hoses.

The transfer piping was custom designed for each instal-

lation to enable minimum transport distances and number

of ow direction changes. Figure 6 shows the pneumatic

transfer piping for the Trenton 5 plant. Trenton 5 was the

only outdoor installation with injection after the existing

ESP into the top ue going to the baghouse. The return ue

form the baghouse is the lower one.

The PAC injection rate was automatically controlled by

the DCS system. The equipment operator could directly

select the required injection rate for unattended operation.

The carbon addition rate is based on a ratio of pounds of

carbon per million actual cubic feet of ue gas at the air

heater outlet (lb/Macf).

Computational Fluid Dynamics (CFD) was utilized in thedesign of individual components as well as to establish injec-

tion locations. The major goals of the CFD analyses were:

• To verify as built piping pressure drop;

• Achieve uniform PAC distribution to each bottle;

• To achieve target distribution of PAC to each lance via

distribution bottles;

• To be able to inject predicted quantities at two elevations

along a given lance; and,

• To achieve desired mixing of PAC in ue gas prior to

particulate collection

Figure 7 shows typical results of PAC particle paths in

the distribution bottle. The biggest issue for the design was

accounting for the high particle momentum. Several bottle

styles were investigated with the bottom fed arrangement

producing the most even distribution. This was also an

advantage for the installation arrangement of the rigid and

exible piping. CFD modeling provides the exibility to

easily customize the distribution bottle design for different

number of lances and preferential distribution as required

for specic applications.

Figure 8 shows the results of PAC ow in an individual

lance. For comparison a photo from wind tunnel testing is

also presented. The CFD model accurately reproduced the

major ow features found in the wind tunnel test. The use

of CFD modeling provides the exibility to easily custom-ize the lance design for multiple/varying levels and varying

amounts per level.

For each plant detailed models of the ue system were

developed including, where appropriate, injection lances,

airheater and particulate collection device. These models

Fig. 3 MercPlus™ injection system equipment.

Fig. 4 MercPlus™ solution injector system.




were used to optimize PAC distribution in the ue gas at the

entry point to the particulate collection devices. A typical

result is shown in Figure 9.

Mercury removal performance testing

Methodology

Mercury concentrations were measured using U.S. EPA

Reference Method 30B “Determination of Total Vapor Phase

Mercury Emissions From Coal-Fired Combustion Sources

Using Carbon Traps”. To determine mercury mass emission

rates and the mercury reduction efciency of the mercury

removal systems, EPA Methods 3 (O2/CO2), 4A (moisture

content), and coal ow were also used (except at Trenton

Unit 5 where coal ow data was not available). The testing

included one-hour sampling runs conducted at the outlet lo-cations. Inlet sampling at Trenton Unit 5 included 20 minute

sampling runs conducted during the one-hour sampling runs

collected at the outlet location. The outlet testing consisted

of sampling of ue gas at the locations described below:

• Lingan Units 1 and 2 - breeching prior to entering the

combined stack

• Lingan Units 3 and 4 - breeching prior to entering the

combined stack

• Trenton Units 5 and 6 – stack; and Point Tupper Unit

2 - stack.

Fuels and PAC properties

Performance testing was conducted during November – December 2009. Colombian coal was red during these

tests. The as-red fuel data are provided in Table 3.

At the time of testing, the coal also had a relatively low

mercury concentration ranging from 0.035 to 0.060 ppm

which is generally more consistent than when blended fuels

are used. The chlorine content of this coal was also low

compared to some of the previous fuels used, ranging from

80 to over 300 ppm, but mostly under 200 ppm.

Fig. 5 Lingan Plant powdered activated carbon silos and feeders. Fig. 6 NSPI Trenton 5, activated carbon distribution piping.

Table 2

Carbon Injection Equipment Selected Design Features

Plant Lingan Point Tupper Trenton 5 Trenton 6

Unit Nameplate Capacity 4 @ 150 MW 1 @ 150 MW 1 @ 150 MW 1 @ 160 MW

Injection Location Economizer outlet Economizer outlet

ESP outlet/

Baghouse Inlet Economizer outlet

# of ues 2 1 1 2

# of PAC distribution bottles 2 1 1 2

# of lances 2 x 6 = 12 1 x 6 = 6 1 x 6 = 6 2 x 5 = 10

# of injection points 24 12 12 20




Two different PAC products were used. For condential-

ity reasons they are described as PAC-1 and PAC-2. ThePAC media is a nely ground, high performance brominated

activated carbon that is specically manufactured for the

removal of both elemental and oxidized mercury from the

ue gases of coal-red boilers. The particle size and den-

sity of this bulk powder are presented in Table 4- Selected

PAC Properties. The particle size distributions of the bulk

powders are presented in Figure 10.

While PAC is a commercially available specialty chemi-

cal, NSPI found that an initial effort was involved in logistics

of co-coordinating fuel burn rate and carbon injection rates

during the testing.

Results and discussion

NSPI conducted guarantee mercury reduction perfor-

mance testing on all seven of its coal-red boilers. Using

the combined technologies of chloride injection to the coal

and PAC injection before the particulate collector, an aver-

age of 70% mercury reduction from the coal potential to

the stack was required. Tests were conducted on all seven

units between November 18 and December 18, 2009. The

purpose of the testing was to verify the percentage of mer-

cury capture at specied MercPlus and PAC injection rates.

Mercury capture was determined by one of two methods:

• On Trenton 5 mercury sampling was performed at theeconomizer inlet to establish untreated mercury emis-

sion rates. These levels were then compared to samples

from the stack.

• On the remaining units, untreated mercury emissions

were calculated from coal composition and coal feeder

data. These levels were then compared to samples from

the stack.

The coal potential mercury emissions ranged from

3.5 – 5.1 lb/TBtu.

The MercPlus solution injection was established based

on augmenting the inherent coal chlorine to reach a target

of 700 ppmdry. For the coal chlorine in the test coals, thechloride addition ranged from 300 – 600 ppmdry. If the

inherent coal chlorine had been high enough, then no solu-

tion would have been required.

PAC injection rates were set to obtain 4 lb/Macf for

Trenton 5 and 6 lb/Macf for the remaining units. The lower

injection rate for Trenton 5 reects the use of a baghouse for

particulate collection, which has much higher contact time

when compared to an electrostatic precipitator, between the

activated carbon and ue gas. Even with 1/3 less PAC, Tren-

ton 5 had the highest capture (~98%) consistent with mercury

Table 3

Test Fuel Properties

Plant Lingan 1 Lingan 2 Lingan 3 Lingan 4 Point

Tupper Trenton 5 Trenton 6

Ultimate Analysis

(% by weight)

C 60.54 59.20 62.29 64.22 63.51 64.81 62.52

S 0.93 0.72 0.97 0.72 .60 0.67 0.72

H2 4.02 3.84 4.2 4.19 4.36 4.04 4.06

N2 1.23 1.21 1.23 1.27 1.35 1.23 1.14

O2 9.22 8.91 9.26 7.82 7.50 3.52 4.50

Moisture 18.34 19.0 15.94 15.7 14.31 14.77 17.74

Ash 5.88 7.12 6.08 6.1 8.38 10.99 9.35

Cl (ppmd) 81.83 81.0 336 169 175 127.5 127.5

Hg(ppmd) 0.065 0.050 0.050 0.058 0.045 0.046 0.050

HHV (Btu/lb) 10,361 10,152 10,834 10,758 11,004 10,506 10,350

Fixed Carbon 44.11 42.52 45.29 45.58 44.26 42.70 41.57

Volatile Matter 31.68 31.36 32.69 32.62 33.05 31.54 31.34

FC/VM 1.39 1.36 1.38 1.40 1.35 1.35 1.33




capture performance on units equipped with a baghouse.

Performance test parameters and results on the aver-

age of several individual test runs are presented in Table

5. The average system-wide capture over the seven plants

of approximately 80% compared well to the performance

target of 70%.

PAC-1 showed a signicant variation in capture efcien-

cies. As it was only used at Lingan, plant design variation

can be ruled out as a contributing factor. It was noted that

the rst two tests (conducted on Lingan 3 and 4) were done

with only a few hours of initial conditioning. It was also

noted that the capture percentage increased as the test period

progressed. Based on this, a minimum 24 hour “seasoning” period was, therefore, established for subsequent tests. Cor-

respondingly, overall capture was signicantly higher on

Lingan 1 and 2 tests. This is indicative of actual performance

on the Lingan units.

PAC 2 was utilized on Trenton 5, Trenton 6 and Point

Tupper. The latter two units are equipped with electrostatic

precipitators and thus can be reasonably compared to each

other. Despite signicant differences in the boiler design,

PAC system and electrostatic precipitators the capture ef-

ciencies are almost equal.

The eetwide mercury reduction measured during the

guarantee testing is summarized in Figure 11. The overall

eetwide mercury reduction of 80% compares well to the

guarantee of 70%.

Mercury removal sorbent costs

Sorbent injection is considered one of the simplest and

most mature technologies for mercury control with relatively

low capital costs (REF 4). However, the operating costs arerelatively high, especially if the system does not include

other air pollutant removal equipment, such as SCR or FGD.

Since the NSPI mercury reduction relies solely on injection

followed by particulate collection, the injection rates and

costs of the mercury control sorbents are relatively high.

For condentiality reasons, NSPI could not include the

actual delivered cost for the mercury sorbents. However,

using a typical basis of $USD1.00 for the cost of calcium

chloride solution per gallon and $USD1.00 for the cost of

PAC per pound the cost for the chloride and carbon sorbents

as a ratio of the plant power output with a range from 1.42

to 2.03 $USD/MW h. When FGD is included the sorbent

requirements are greatly reduced and operating costs aresubstantially less. For example, The Newmont Mining TS

power plant utilizes a spray dry/baghouse FGD system that

operates with only 0.3 lb/Macf Br-PAC and 100 ppm CaCl2

for an average cost of $0.13/hr (REF 5).

Fig. 7 CFD simulation of PAC flow in the distribution bottle.

Table 4

Selected PAC Properties

PAC Product Mean Particle

Size (microns)

Bulk Density

(lb/ft 3 )

PAC 1 12.14 30.1

PAC 2 17.82 21.6

Fig. 8 Comparison of CFD simulation and wind tunnel testing of

individual lances.

Fig. 9 Lingan Plant CFD for PAC distribution at the ESP inlet.




The majority of the cost is for the brominated PAC,

which averaged 89%. Depending on the overall boiler

eet utilization, and coal chlorine content, the total annual

cost for mercury sorbents could range from about $USD

10,000,000 – 15,000,000 annually based on this available

commercial information.

Conclusions

The installation of seven mercury abatement systems

was successfully installed and tested in 2009 at Nova Scotia

Power. These projects were initiated in response to the pend-

ing provincial annual cap not to exceed 65 kg of mercury.

The two components of these installations were enhanced

mercury oxidation using calcium chloride application to the

fuel feedstock in combination with a brominated PAC. An

interesting feature of the systems installed on the six units

without a baghouse is the novel hot-side PAC injection lance

system located before the air preheater.

Following construction, tests completed at all unitsyielded results that exceeded the targeted 70% mercury

capture and positioned NSPI to meet the provincial regula-

tion in 2010.

References

1. Canadian Electricity Association (CEA) Mercury

Program. Retrieved June 13, 2010, from http://www.

ceamercuryprogram.ca/EN/program_overview.html

2. Canada-Wide Standards for Mercury Emissions from

Coal-Fired Electric Power Generation Plants, Endorsed by CCME Council of Ministers, October 11, 2006

3. Schedule C of the Air Quality Regulations. Environ-

ment Act of Nova Scotia. Amended September 25, 2007

4. Institute of Clean Air Companies, Sorbent Injection

Technology for Control of Mercury Emissions From

Coal-Fired Boilers, ICAC Fact Sheet 051506.pdf

5. Jankura, B.J, et.al, Cost Effective Mercury Emissions

Control At The Newmont TS Power Plant, Electric

Power May, 2010, Rosemont, Ill

Key words

Calcium Chloride

Powdered Activated Carbon

Mercury Control

Nova Scotia Power

The Babcock & Wilcox Company

Fig. 10 PAC particle size distribution. Fig. 11 NSPI fleet-wide mercury reduction.

Tables 5

NSPI Plant Mercury Removal Performance

Plant Particulate

Control

Cl

Addition

ppmd

PAC

Addition

lb/Macf

PAC

Type

Boiler

Load

MWg

Coal

Cl

ug/g

Coal

Hg

ng/g

APH

Tin

F

APH

Tout

F

Coal

Hg

g/hr

Stack

HgT

M30B

g/hr

Stack

HgT

M30B

lb/TBtu

Hg

Removal

Eff %

Lingan 1 ESP 600 6 PAC-1 161 82 65 N/A 331 2.61 0.27 0.31 92

Lingan 2 ESP 600 6 PAC-1 149 81 50 745 319 2.54 0.67 0.95 75

Lingan 3 ESP 300 6 PAC-1 165 336 50 713 340 3.12 1.09 1.85 59

Lingan 4 ESP 300 6 PAC-1 162 269 58 713 340 3.36 1.26 2.10 58

Point Tupper ESP 500 6 PAC-2 157 175 45 672 329 2.76 0.41 0.58 84

Trenton 5 BGH 550 4 PAC-2 118 127 46 N/A 302 3.34 0.04 0.04 99

Trenton 6 ESP 550 6 PAC-2 158 127 50 703 267 2.91 0.38 0.60 86



Copyright © 2010 by Babcock & Wilcox Power Generaon Group, Inc.

a Babcock & Wilcox company

All rights reserved.

No part of this work may be published, translated or reproduced in any form or by any means, or incorporated

into any informaon retrieval system, without the wrien permission of the copyright holder. Permission re-

quests should be addressed to: Markeng Communicaons, Babcock & Wilcox Power Generaon Group, P.O.

Box 351, Barberton, Ohio, U.S.A. 44203-0351. Or, contact us from our Web site at www.babcock.com.

Disclaimer

Although the informaon presented in this work is believed to be reliable, this work is published with the

understanding that Babcock & Wilcox Power Generaon Group, Inc. and the authors are supplying general

informaon and are not aempng to render or provide engineering or professional services. Neither Babcock

& Wilcox Power Generaon Group, Inc. nor any of its employees make any warranty, guarantee, or representa-

on, whether expressed or implied, with respect to the accuracy, completeness or usefulness of any informa-

on, product, process or apparatus discussed in this work; and neither Babcock & Wilcox Power Generaon

Group, Inc. nor any of its employees shall be liable for any losses or damages with respect to or resulng from

the use of, or the inability to use, any informaon, product, process or apparatus discussed in this work.

Documents

BR 1844 Babcock Mercury