Snowbird Jr m

8/10/2019 Snowbird Jr m

1/53

Mooney Session 4 Coal Blending

Fuel Quality Conference 2006 page 1-1

Coal Blending for Operational ExcellenceJames R. Mooney

45 Saint Andrews DriveBeaver Falls, Pa 15010


2/53



ABSTRACT

This paper will review the experiences of the author during and following the

installation of a Digital Fuel Tracking System (DFTS), supporting initiatives to reduce

fuel cost by blending of opportunity fuels, increasing fuel flexibility, and as experience

was gained to address issues with auxiliary equipment that went far beyond the original

expectations for operating with precision. Digital Fuel Tracking is not artificial

intelligence; instead it is a system of compiling all available information from existing

manual records and various software systems installed, making it accessible to plant fuel

handling, operations, and engineers. These people use the information to solve problems

in pursuit of lower fuel costs and higher plant capacity factors.

Operational excellence targets top decile capacity factors (+84%

capacity factor for base load units)

Fuel Flexibility is a program of purchasing least cost BTUs to

accomplish a plants mission

Attempting to meet increased output while simultaneously pursuing lower cost

BTUs or fuels of opportunity require accuracy in fuel blending to accomplish the plant

mission. The increased complexity of power plants through the addition of FGD, SCR,

and waste disposal systems require the fuel recipes be as accurate as possible.

Keywords: Fuel, flexibility, operational, excellence, digital, tracking


3/53



I. Introduction

The Digital Fuel Tracking System (DFTS) brought together information from:

Corporate Fuel Management System (accounting system used to

purchase, track and pay for fuel)

Yard Handling manual records

OSI Pi Software inputs from field measurement devices

Through fuel testing in 1997 the plant was keenly aware of the influence

coal quality was having on plant performance since at least one unit would experience a

forced outage each year due to slag falling and creating tube leaks. A fuel calculator

(figure 1) was developed at the plant and modified with predictions for slag based on fuel

characterization investigations into an incident of slag falling in 2000. Elimination of

these incidents was a top priority, and controlling the blends of coal, quickly changing

when necessary, was the solution for this particular problem.

With experience at tracking the coal quality it was found many other derating

occurrences of the plant were fuel related and by knowing the coal quality currently

burned it was possible to recognize negative conditions and create a positive outcome

through changing the fuel quality. In 2004, the Mansfield Plant achieved top decile

performance. First Energy Baseload units capacity factors reportedly averaged 69%

from 00-02, and 04-05 were 85%, with Mansfield at 97.5% CF through the first half of

2005.


4/53



II. DISCUSSION

In the 1990s significant changes in preparation for the onslaught of deregulation

were descending on the plant. A number of initiatives were developed to position the

company for the future deregulated environment; two of these were developed in

different areas of the corporation, results dependant on capability to perform. Initiative

one was Fuel Flexibility, deemed important to my plants going forward cost because the

long term contract price for coal was significantly above the depressed spot market coal.

An opportunity existed to buy over 2.0 million tons of higher ash, lower BTU coal, if the

plant could successfully burn it.

The second initiative was to Operate with Precision. In 1997 I was offered the challenge

to first look at various ways to reduce the going forward fuel and lime costs at the plant.

It was after accepting this position I became aware of the second initiative to operate with

precision, defined as being available for full load, eliminating operator errors, improving

Forced Outage Rates and eliminating environmental excursions, spills etc.

Experimenting with changing coal did not seem like a particularly good idea in light of

the Operate with Precision challenges, but it helped justify the installation of the Digital

Fuel Tracking System.

In 1997 almost all of the coal delivered to the plant was by 1500 ton river barges,

under one contract. This long term contract ran from 1980, and was to expire or be

renegotiated by end of 1999. The quality was on a monthly average basis, with premiums


5/53



paid for better than average quality, and penalties should the quality be worse than

specified up to a maximum or reject level for BTU/Ash/Sulfur.

In reality the coal is:

Better than the contract specification

Actually hits the specification quality as delivered (rare)

Occasionally is significantly worse than specified

The variability of the delivered coal is what makes life interesting in the plant.

Vendors supplying coal on monthly average have financial incentives to deliver the better

than specified coal and using a monthly average provides them some room to react whenwhat comes out of the ground is significantly worse than specified. The power plant

needed to react in a similar fashion when the coal delivered quality went from better than

specified to significantly worse, and the solution was to install a Digital Fuel Tracking

System with shared access across the company, and near real time quality information for

as delivered coal.

A Fuel Flex Team of employees made up of corporate and plant personnel

included the fuel purchasing department, the plant yard, operations and technical

services, and were led by conversion economics. The type of information needed was

defined by the Fuel Flex team as a Digital Fuel Tracking System capable of:

Recording and historize field inputs (scales, instrumentation, belt speed) to

track fuel to specific locations within plant or yard storage

Accessible on company LAN

Customizable reports by date/time for delivery, unloading, bunkering,

quality, and quantity (samples, figure 2)


6/53



OSI Pi monitoring capability

Minimization of manual recordkeeping

III. RESULTS

The digital fuel quality system received information from various sources

including Fuel Management System data, plant records, and field inputs to

accurately track the coal quality and quantity into each pile and the bunkers. In

addition a pile was created that tracked the floating inventory, at the time over

95% of the coal was received by barge and the remainder was truck delivered.

Truck scales were also tied into the DFTS system for recording deliveries, and

this was typically into the high ash pile. The petcoke was limited to 20% by

weight of the total consumption by the state department of environmental

protection, and was therefore stored and loaded onto belts separately for precision

in blending and control.

To address accessibility the Digital Fuel Tracking System information was

available to all company personnel via the LAN system. Customized reports

(samples attached) were created by the developers, Engineering Consultants

Group-Inc, for monitoring deliveries, bunker loadings, quality, quantity, and

analysis with costs for each of the fuels delivered. DFTS screens were developed

within the OSI Pi screens, mimicking the plant control screens for layouts and

included inventory information where anyone with access to OSI Pi software

could call up and view inventory by pile including the mathematical calculation of

on-the-belt specifications for coal going to the bunkers.


7/53



The ability to customize the reports for the various groups using them was

important to the fuel flex team. The diversity of the team selection and proactive support

from corporate leadership insured purchasing departments, plants, transportation issues

that were all of different levels of importance to the missions of each group were

addressed by customization of the reports. Transportation was primarily interested in

how many barges were being emptied for a given period to measure carrier performance

contract items, and could request just the information that was needed for their job

without additional items like quantity unloaded that was of importance to other groups.

The Digital Fuel Tracking System became the tool that tied all of the other availableresources together from the Fuel Management System (accounting and contract

monitoring software) to OSI Pi, and manual entry logs in a one stop customized format.

OSI Pi was already available on the Corporate LAN, making it the logical location for

data storage.

Manual data entry reports from bunker loadings were a frequent source of

error, which were identified as a significant source of inventory discrepancy and

adjustments at annual flyovers. One of the goals for DFTS was elimination of as much

manual data input as possible and reliance instead on the instrumentation, scales and field

inputs to update records. This initiative required that the OSI Pi system reliability be

increased to update reports, and a monitoring system was installed that would catch and

report any gaps in provided information, flagging appropriate personnel to a problem

with the information gathering and the need to perform manual calculations. As a result

of the increased monitoring OSI Pi availability had increased to +99%, and daily reports

are updated automatically for reporting at daily tech team meetings.


8/53



As an update to the paper titled Development of Fuel Quality Predictor

delivered at the last Engineering Foundation Conference by M.Valach, et al the plant did

in fact install leading edge sootblowers on the Mansfield units discussed in the paper, and

combined with fuel monitoring has avoided any new incidents of slag fall outages since

the 2000 incident reported on at the 2001 conference on Fuel Quality Impacts. The

ability to eliminate these forced outages by controlling the blends of fuels at the plant

proved the value of digital fuel tracking as a tool for increased capacity out of the plant.

The Fuel Flex initiative also supported the installation of rapid discharge rail unloading

equipment at Mansfield and other plants to increase access to other coal supplies,enabling security of delivery and increased blending capabilities.

As new systems were installed for NOx and Opacity Mitigation requiring

new commodities the DFTS was upgraded to include quality and quantity monitoring tied

to the OSI Pi software. DFTS was upgraded to include accurate tracking of aqua

ammonia and sodium sulfite, and the use of bar coding with scanners was instituted to

automatically scan the vendor quality information, with reporting functions similar to

coal and automated recording of delivery and unloading times.

In addition to identifying and controlling the slag from the fuel burned coal

tracking revealed solutions to derate in other areas of interest to the technical group.

Some of these were:

Air Quality Control System chemistry

Gypsum Production

NOx

Reagent consumption


9/53



Lime area operations

Fuel Inventory Control

As experience with control of the fuel blends to manage inventory or combustion

brought discussions of the blends loaded to the bunkers to the daily technical team

meetings, other areas of the plant took note of the changes, good and bad, that caused

requests for fuel blends from unexpected areas.

Air Quality Control System

It was obvious and well known that the scrubbers were installed at the plant for

the control of SO2 emissions. What was surprising to many of the technical supervisorswas how varied the coal sulfur content was, and that many times they were searching for

equipment malfunctions to explain sudden changes in scrubbing capability prior to the

DFTS and fuel quality discussions they were unaware of how much the coal quality was

changing. A unit may have been happily running along for 6 months receiving a coal

with sulfur content of 3.0 #/MMBTU, only to have the mine hit a pocket of 3.75

#S/MMBTU, causing issues with the ability to transport enough lime to stay in

compliance, problems with slaking equipment keeping up with the higher demand for

lime, discovery that piping that could easily transport enough lime for the lower sulfur

content fuel was partially plugged and looking for the reasons as to why the piping

suddenly plugged, when what actually suddenly changed was the fuel quality. By

tracking the fuel quality into the yard storage it was possible to address issues like this

using precision fuel blending selecting from the lower sulfur content fuel to control the

emissions and avoid the derate.


10/53



Gypsum Production

A new facility was constructed to produce gypsum at the plant, called the FOG

system for Forced Oxidation Gypsum. The plant produced high quality gypsum for a

customer to produce wallboard, and the project was very important to both companies

since it was part of the life extension plan for sludge disposal for the power company, and

the gypsum was to be the main ingredient of wallboard for the plants customer. The

original fuel flex team never anticipated requests for specific fuel would be coming from

the FOG plant when the DFTS was installed, but with minimum contract supply

agreements to the customer the plant was thrust into the position of commodity supplier,with penalties for not meeting monthly and annual targets. One of the units produced

gypsum more efficiently than the others, and was targeted to get higher sulfur content

fuels as a result. Targets were discussed daily and fuel blends adjusted accordingly to

meet the contract minimums.

NOX Control

Combustion engineers responsible for NOx monitoring requested seasonally

adjusted blends of fuel for the NOx season of May September. Consideration was

given to the impact of fuel blends on SCRs and opacity mitigation systems, reagent

consumption, and had the effect of burning more opportunity fuel out of season to save

certain better performing fuels for the critically important NOx season.

Reagent Consumption

Tracking aqua ammonia and sodium sulfite through the DFTS provided similar

benefits to monitoring coal quality. Budget costs for the reagents could be affected by

the coal quality, and blending the coal to produce the overall lowest cost for the plant was


11/53



enabled by knowing what we were consuming from actual field inputs. Quality control

was improved as a result of tracking the vendor delivering reagent directly into Pi.

Lime Area Operations

Lime area operations are responsible for delivery of lime slurry at specific

density and temperatures to effectively scrub SO2. Prior to DFTS the operators

frequently placed additional equipment in service to provide response time to control the

occasional spikes that put the unit out of compliance. With better blending control at the

plant the additional equipment was no longer required because advance warning of sulfur

changes in the coal quality was available.Fuel Inventory Control

Digital Fuel Tracking improved the control of fuel inventory in three

significant ways:

1. By placing the coal into specific piles by quality the annual flyover for

inventory adjustment was made more accurate as to the value of the coal

in each pile

2. The variance report of DFTS comparison with vendor supplied

information for vendor weights provided alarming to the appropriate

employees whenever a variance of more than a set amount, typically >4%,

in the two weights appeared, prompting an immediate investigation to

identify the reason.

3. The DFTS daily numbers from field input devices replaced manual logs

and bunker reports, which were then typed in to the Fuel Management

report. Transcription errors, which unit the coal was delivered to, and


12/53



what the quality was were all found to be frequently in error and the cause

of inventory adjustments easily reaching into the millions of dollars

following flyovers to adjust inventory. DFTS accuracy was a profound

improvement.

CONCLUSION

In 2003 the Mansfield Plant, the largest plant in the First Energy system, set

a new generation record of 17.4 million megawatt hours. In 2004 the plant broke this

record, exceeding 18 million megawatt hours, and in June of 2005 was on course to re-setthe record book again with a reported capacity factor of 97.5% for the first half of 2005.

First Energy, in response to a question during the 2 nd quarter 2006

performance web cast to investors, reported the fleet fossil generation capacity factor was

again top decile at 88.9%, and the company is on target to once again break their

generation record.

Other significant indicators of plant performance include:

2000 through 2003 maintained a 2.0% Plant forced outage rate.

An industry benchmark of 327 days continuous operation set by Unit 3 in

2001.

In 2003 the Mansfield plant forced outage rate was only 1.14%!

Accomplishing all of the above required people, technology, and equipment to

achieve. The people portion is extremely important to these accomplishments and

included commitments from the corporate leadership down through the newest yard

operator pushing in coal, and frequent discussion with fuel purchasing department to


13/53



qualifying new fuels. With Digital Fuel Tracking in place identifying root cause of issues

with auxiliary systems becomes possible by utilizing OSI Pi to identify when the coal

quality change occurred and the effect when a blend change is made to address the

problem. This brings theory and reality in close proximity, the more precision used in the

blending the better characterized the impact of the fuel on the specific unit it is burned

on. As a minimum field inputs to an OSI Pi or similar system can provide significant

benefits to a plant in communicating what fuels are available, and in more sophisticated

systems such as Detroit Edisons Monroe station providing an early warning screen based

on output of an On-Line Analyzer, that is continuously advising the Control RoomOperators and Shift Supervisors of the quality of the coal with recommendations of

courses of action, is a significant upgrade.

Utilities seeking to increase plant output while simultaneously seeking lower cost

fuel will benefit by putting all of the tools together in a Digital Fuel Tracking System that

enables people to use the fuel available to them to achieve the best possible output. A

cheap coal that causes a unit derate for slag or emissions is not a good value.

REFERENCES:

Development of Fuel Quality Predictor for Three 900 MW e Units, Engineering

Foundation Conference, Snowbird, UT, 2001, M. Valach, et al


14/53



Figure 1, Coal Calculator


15/53



Sample Report Tables, courtesy Engineering Consultants Group, ECG-Inc, Akron, Oh


16/53




17/53

Mueller Session 4 Intelligent Cleaning


Intelligent on-load cleaning technology to increase power boilerperformance

Franz Bartels, Stephan Simon, Manfred Frach and Christian Mueller

Clyde Bergemann GmbH, Schillwiese 20, D-46485 Wesel, Germany

Abstract

Slagging and fouling are the most common reasons for reduced boiler availability andefficiency. The number of unscheduled shutdowns as well as the degree of efficiency loss isstrongly dependent on the fuel quality and increase with the use of low rank fuels and fuelmixtures. The most efficient measure to increase boiler efficiency and availability is theselective on-load cleaning of different boiler sections with appropriate cleaning equipmentcontrolled by intelligent diagnostic systems.

In this paper an efficient on-load cleaning system is presented, combining water cannontechnology with advanced heat flux measurements to a closed loop furnace optimizationsystem for detecting and cleaning boiler regions with unacceptable high ash deposition. Themain focus is put on the optimized selective cleaning of the furnace region to achieve a lowfurnace exit gas temperature since the latter is a direct measure for the evaporator efficiencyand an indicator for fouling on the heat exchanger surfaces in the convective pass.

The on-load boiler optimization system was applied to a 600 MWe utility boiler fired withlignite and bituminous coal. Already during the first weeks of operation with the optimizationsystem a significant improvement of the boiler efficiency was observed. The average heattransfer from the furnace to the steam cycle increased by 30% and the furnace exit gastemperature decreased by about 80 degrees.

Keywords On-load boiler cleaning, intelligent sootblowing, water cannon technology,heat flux sensors

Corresponding author: [email protected]


18/53



1. Introduction

Liberalisation of the energy market was a political milestone for Europe with quite differentimpacts on energy consumers and producers. On the one hand the liberalisation led to areduction of electricity prices due to enhanced competition. On the other hand it increased the

pressure on the power producers to save production costs and increase production efficiency.Since the fuel price is a considerable fraction of the production costs, nowadays, whenever

possible, low rank coals from the international market are fired. However, these import coalsshow frequently varying fuel characteristics and may have a strong tendency for slagging andfouling. As a result of non-uniform slagging on the evaporator walls, boiler availability issignificantly reduced, and so is the boiler efficiency. Furthermore, the reduced heatabsorption in the radiative section of the boiler leads to an increased furnace exit gastemperature (FEGT) which may again lead to enhanced fouling in the convective section ofthe boiler.

However, not only imported coals may cause serious operational challenges. For the lignite-fired utility boilers in the Rheinish region between Cologne and Aachen in Germany a severefuel change took place in recent years. Opening of the new Hambach open-cast mine, whichin the years to come will supply a considerable fraction of the lignite for the installed power

plant capacity of overall some 10.000 MW in the Rheinish lignite region, entails a significantchange of the fuel properties [1]. The lignite from the Hambach open-cast mine ischaracterised by a particularly high calorific value and a high iron and sodium content(Tab.1). The higher calorific value leads to higher combustion temperatures and, as a result,to an increased furnace exit gas temperature. The increased iron content and thealkaline/silicon ratio lead to a lower ash fusion temperature and enhanced fouling,respectively [2]. These two processes, especially when occurring together, cause considerableslagging in the furnace (Fig. 1) and molten ash in the convective heat exchanger region.Efforts to control this behaviour by means of operational changes were not satisfactory.

Figure 1 Slagging around coal burner mouth.

To limit the rising FEGT it is therefore necessary to increase the heat transfer to the radiativeheat exchanger surfaces in the furnace. Hence intelligent high performance on-load cleaning


19/53



of the furnace walls is a precondition for save and efficient operation of a steam generatorfired with demanding fuel or fuel mixtures.

Table 1 Fuel analysis Hambach coal.

2. Optimised furnace cleaning in a 600 MWe unit

The concept for an intelligent high performance on-load cleaning system will be introducedand described in detail for the steam generator of a plant fired with lignite and bituminous

coal. The steam generator is a tower type boiler with a capacity of 600 MWe, was taken intoservice in 1974 and has a membrane wall design with a cross section of 20 x 20 meters(Fig. 2). Over the years several modifications have been made to this boiler in order tooptimise the combustion conditions and to reduce emissions.

Figure 2 Steam generator.

Data for Rough Coal from Open Mine Hambach

Average Coal Characteristics

Sulphur CalciumPotassiumSodiumIron[ppm]

Ash[%]

Calorific Value[kJ/kg]

0,26550015070027003,0 to 3,59700

[ppm] [ppm] [ppm] [ppm]


20/53



2.1 Cleaning equipment

The conventional method to clean furnace walls from slagging is the use of wall deslaggers.These cleaning instruments are soot blowers with two opposing nozzles that enter the furnacewith a rotating motion while blowing the steam slightly inclined backwards to clean thefurnace wall behind. The result of wall deslagger operation is patterns of clean circles on thewall. However, in between these circles uncleaned regions remain (Fig. 3). The sum ofuncleaned surface areas with a significantly reduced heat transfer rate amounts to up to 35%.

Figure 3 Cleaning regions for conventional wall deslaggers vs. water cannons.

A novel instrument for furnace cleaning is water cannons (WLB). The water cannon

technology (Fig. 4a) is characterised by a concentrated water jet throughout the furnace,impinging on the opposite wall. Due to a controlled motion of the cannon, the impinging jetdescribes a meander-shaped cleaning pattern on the wall surface that is set up individuallyaccording to the cleaning needs (Fig. 4b).

Figure 4 a) Water cannon [5]. b) Cleaning pattern.

1 0 9 0

1 0 9 0

1 0 9 0

1 0 9 0

BlowingJet

InstallationPlace

BlowingJet

InstallationPlace


21/53



The cleaning mechanism of water cannon basically differs from that of a wall deslagger.While the wall deslagger operates with high pressure steam as cleaning medium, watercannons operate with water that is injected through high performance nozzles into thefurnace. While impinging on the opposite wall the produced spray of droplets penetrates the

pores of the deposits. Inside the pores the water suddenly evaporates and the volume rapidlyexpands. This expansion loosens the deposits from the heating surface and removes it (Fig.

5). The overall efficiency of this cleaning process depends on the optimum penetration of thedeposit by water and is influenced by:

- Velocity and impact angle of water jet at arrival on the opposite wall

- Impinging water quantity

- Characteristic of the deposit surface (porous vs. glassy)

Figure 5 Cleaning concept of water spraying.

Compared to conventional wall deslaggers the use of water cannon technology has significantadvantages for the power plant efficiency:

- Increase of heat absorption in the evaporator and significant reduction of furnace exit gastemperature

- Reduced operational costs, since water is used as cleaning medium instead of superheatedsteam

- Reduced maintenance costs since one water cannon replaces up to 15 conventional wall

deslaggersThe common criticism that water jets impinging on furnace walls may enhance the risk ofthermal impact and the related reduction of life time for the wall tubes of a steam generator isincorrect. Operational experiences since more than 12 years have shown that the risk ofthermal impact can be excluded if the cleaning parameters for water cannon are correctlyadjusted. This has also been shown by extensive measurements and calculations performed

by others [3].

T u b e

W a

l l W a

t e r

/ S t e a m

T u b e

W a

l l W a

t e r

/ S t e a m

T u b e

W a

l l W a

t e r

/ S t e a m


22/53



2.2 Local detection of slagging on furnace walls

Due to the typical non-uniform deposition pattern on furnace walls it is highly inefficient toclean the entire furnace anytime a cleaning sequence is started. Hence an on-load detectionmethod is required that allows to identify the furnace wall regions that are covered withdeposits above an acceptable upper limit. The detection method of choice is a novel heat fluxsensor based system further developed from earlier work by Davidson [4] and especificallydesigned for this application (Fig. 6).

Figure 6 Heat flux sensor integrated into an evaporator tube.

For installing the sensor an approximately 400 mm long evaporator tube section is replacedwith an individually manufactured tube sensor with the following design features:

- Two redundant thermocouple pairs on the tube side facing the furnace

- Circular tube surface is maintained

- Thermocouples arranged such that no protruding edges can enhance deposit formation

- Water/steam flowing through the sensor tube section

- Thermocouples inserted into the tube wall ensuring a controlled low pressure drop on thewater/steam flow inside the tube

- All four measuring sections resulting from the redundant arrangement of the thermocouplescalibrated individually for each boiler

- Heat flux sensor designed and introduced such that only vertically incident radiation ismeasured

Due to the special design the heat flux sensor is a reliable instrument and has shown verylong lifetimes of more than 5 years.

Combination of the heat flux sensor with an optimisation system that correlates thethermocouple measurements delivers the respective local heat flux value in kW/m.For the evaluation the heat flux data and the decision on the respective cleaning actions notonly the absolute value of the heat flux is relevant but also the characteristic pattern of theheat flux signal over the operational time (Figure 7):


23/53



Figure 7 Time history of measured heatflux.

Period a) For a clean heating surface a high frequency of the signal is measured. Acontinuous decrease of the absolute heat flux value indicates the beginning ofdeposition. While the heating surface with a first layer of deposits catches moreand more ash particles, the heat flux decreases continuously and consequently theabsolute temperature at the deposit surface rises.

Period b) When the surface temperature reaches the sintering temperature (critical point) theash particles adhere to each other and the heat flux decreases suddenly and fast.

Period c) During this period the measuring signal of the heat flux is considerably moredamped because of the slag formation, but the heat flux still decreasescontinuously.

Experience has shown that the optimum moment to start cleaning with water cannons is the beginning of period b).

2.3 Intelligent on-load cleaning concept

The overall concept of an intelligent on-load cleaning system is sketched in Figure 8. Theheat flux sensors are installed at preselected, optimum locations on the evaporator walls. Themeasuring signal of the sensors is transmitted through a multiplexer to the data acquisitionsystem where it is complemented with available plant data such as boiler load, coal milloperation, etc. The optimisation system is monitoring the situation online and is calculatingand analysing the current slagging situation of the furnace. In parallel, the system is doing the

optimisation calculation following conditions such as:- Optimum point in time for starting the cleaning action

- Selection of required cleaning zone

- Selection of optimum parameters for the individual cleaning zones (e.g. jet velocity, waterquantity, cleaning pattern)

0 1 2 4 5 3

Time [ h ]

100

80

60

40

20

0

Heat Flux [ kW/m 2 ]

a cb

Critical Point

a cb

Critical Point


24/53



Figure 8 Furnace optimization system.

As a result, the optimisation system determines the required cleaning actions and controls thePLC based water cannons.

3. Results

The investigated boiler is equipped with 14 new water cannons replacing 72 conventionalwall deslaggers. The water cannons are installed at three levels, each equipped with fourcannons, and 2 additional cannons in the bottom region, for cleaning the ash hopper section(Fig. 9).

Figure 9 Positioning of water cannons in the boiler.

SFX-Senso rs

Furnace CleaningWhere?When?How?

Eng ineering Too lVisualisation

PL C

SFX-Senso rs



SFX-Senso rs



PL C

Data Aquisition/Optimisation

System

Heat fluxsensors

SFX-Senso rs



PL C

SFX-Senso rs



SFX-Senso rs



PL C

Data Aquisition/Optimisation

System

Heat fluxsensors

4 WLBs+ 62,500 m

4 WLBs + 48,743 m

4 WLBs+ 36,416 m

Flue GasRecirculationDuct

Burner

4 WLBs+ 19,371 m

20 m 20 m

20 m 20m

4 WL Bs+ 62,500 m

4 WLBs + 48,743 m

4 WLBs+ 36,416 m

Flue GasRecirculationDuct

Burner

4 WLBs+ 19,371 m

20 m 20 m

20 m 20m


25/53



The diagnostic equipment of the on-load cleaning system is based on 22 heat flux sensors thathave been installed during a short shutdown of the boiler (Figure 10). The sensors and theoptimisation system are configured according to the overall requirements of the steamgenerator.

Figure 10 Positioning of heat flux sensors and water cannons.

Since the introduction of the fully automatised intelligent on-load boiler cleaning conceptindicated a significant change in operation, an introduction strategy with different phases waschosen, starting from manual via the first automatised to the final automatised and optimisedoperation.

During the first months of operation with the new concept, at the request of the operator, themeasured data of the optimisation system were only collected, monitored and evaluated. Thewater cannons were not yet controlled automatically but by default operated once peroperating shift (8 h). The boiler measurements and the cleaning recommendations of theoptimisation system in these first weeks already showed that it was quite sufficient to activatethe water cannons at the top level (area of flue gas recirculation ducts) only once or twice aweek, meaning that these areas were cleaned far too frequently in the past. On the other hand,the collected data also showed that in the burner belt area a water cannon cleaning frequencyof two or three times per operating shift (8 h) was necessary, depending on the coal qualityfired. This means that cleaning in that particular area of the furnace was far too seldom in the

past. As a result, deposits on those walls got sintered or even molten and successful cleaningwas not achieved any longer since the water was not able to penetrate the deposit.Furthermore, during this initial phase it became obvious that deposition on the variousfurnace walls changed depending on the coal mill operation.

After a few weeks of data collection, a manual interface was introduced to transfer thecleaning recommendations of the optimisation system to the water cannon control system andto start selective cleaning of specific furnace zones. Figure 11 shows the average heatabsorption of the evaporator. This value is calculated by integration of the 22 heat fluxmeasurements and therefore reflects the measured effectiveness of the evaporator.

Right Sidewall Front Wall Left Sidewall R ea r Wall+ 62.500m

+ 48.743m

+ 36.416m

+ 19.371m

WL B

WL B

WL B

WL B

Right Sidewall Front Wall Left Sidewall R ea r Wall+ 62.500m

+ 48.743m

+ 36.416m

+ 19.371m

WL B

WL B

WL B

WL B

Heat flux sensor

Water cannon

Heat flux sensor

Water cannon


26/53



Figure 11 Heat transfer of evaporator walls averaged over 22 heat flux measurements.

The diagram shows an increase of 25 kW/m 2 for the average heat flux of the evaporator. Theincrease refers to the former situation where the water cannons have been operated only once

per shift (8h). The 30% higher heat absorption leads to a reduction of the furnace exit gastemperature (FEGT) by 78 K for this steam generator. This significant drop in furnace exitgas temperature may lead to reduced deposit formation on the first heat exchangers of theconvection pass since the reduction in FEGT lowers the ash particle temperature.Furthermore, the reduced temperature level of the fluegas leads to a reduction of the sprayingrates for the reheater and results in a lower boiler outlet gas temperature, which implies

reduced waste gas losses. These two effects may lead to an increase in boiler efficiency of upto 0.6 %.

After eight months of operation the fully automatised mode was started. In this phase theoptimization system forms a closed loop, controlling the selective cleaning actions in thefurnace with the appropriate measures when and where required.

After total operation of 12 months the collected data and operating experience indicated theneed for more detailed information at selected regions of the burner belt. Therefore sixadditional heat flux sensors were installed and integrated into the optimization system.Further evaluation of the intelligent on-load boiler cleaning system now equipped with 28sensors showed a reliable monitoring of the slagging situation in the furnace.

However, improvement of the cleaning efficiency was still possible by minor adjustments tothe water cannon control and the optimization of the cannon nozzle geometry which resultedin a more targeted water jet.

4. Conclusions

Ash deposition in utility boilers increases with the use of low rank fuels and fuel mixtures.Above a boiler specific critical deposition rate the overall availability and efficiency of theunit may decrease significantly. A first measure to control the overall slagging and fouling

Figure 11: Average Heat Flux of Evaporator - RWE Rheinbraun PS Niederauem Unit G06.09.00 09.09.00 12.09.00 15.09.00 18.09.00 21.09.00 24.09.00 27.09.000

20

40

60

80

100

120

140

Time

14.09.2000

Know-How-Transfer and

Training of operators

0

20

40

60

80

100

120

140

14.09.2000

Know -How -Transfer and

Training of operators

A v e r a g e

H e a

t F l u x

[ k W / m

2 ]

Start ofoptimised

selectivecleaning


27/53



situation in a boiler is the optimized and efficient cleaning of the heat exchanger surfaces.Hereby the furnace exit gas temperature can be controlled and kept below the fuel specificash melting temperature. Furthermore, the heat transfer rate of the evaporator will beincreased and so will be the efficiency of the steam generator.

The presented work deals with the installation of an intelligent on-load boiler cleaning systemto a 600 MWe steam generator fired with lignite and bituminous coal. In this closed loopoptimization system water cannons are combined with advanced heat flux sensors toefficiently clean the furnace depending on the current operational condition of the boiler.

Operational experience over several months indicated that the presented furnace optimizationsystem enables to minimize furnace slagging and superheater fouling significantly, to reducewaste gas losses and to increase the overall boiler availability and efficiency.

5. Literature

[1] Glaser, W.; Kulik, L.; Neuroth, M.: New Findings regarding slagging behavior ofdifferent qualities of lignite from Rheinish region, VGB PowerTech 10/2002, pp. 100-107.

[2] Couch, G.: Understanding slagging and fouling in pf combustion , IEA Coal Research,Vol. 72, p. 48, 1994.

[3] N.N., Cleaning with the new water lance blower WLB 30 , Experience Report RWEEnergie AG, Eschweiler, 1995.

[4] Davidson, I. S.: A practical on-line solution to control ash deposition , Conf. on theimpact of ash deposition on coal-fired plants, Taylor & Francis, pp. 693-702, 1994.

[5] N.N., www.clydebergemann.de , 2006.


28/53

Batton Session 4 Oxistop Coatings


A Technical Paper:

Oxistop CoatingsFor Boiler Tube Maintenance

Prepared for the:

TECHNICAL CONFERENCE

Impacts of Fuel Qualityon Power Production

October 29 November 3, 2006

Snowbird, Utah

T. E. BATTON - PRESIDENT

OXISTOP, LLC, 6331 MARKET STREET, BOARDMAN, OHIO USA 44512


29/53



Introduction

OXISTOP, LLC, an Ohio based Limited Liability Company, is the sole distributor andapplicator of the Oxistop line of ceramic coatings developed specifically for the fossil fuel

power generation and traditional refractory markets. Both markets currently experience failureand high maintenance costs related to maintaining key components of production.

Power generated from fossil fuel combustion relies on boiler tube integrity. Boiler tube failurecan be related to the destructive nature of slag, oxidation, corrosion, erosion and abrasion.These issues are becoming increasingly important due to the worldwide demand for more

power and operational controls related to emissions regulations.

At a fraction of the cost of current alternatives, the Oxistop line of coating materials has been

proven to stop or minimize the damage caused by these destructive conditions. As an added benefit of utilizing these materials, boiler tubes coated with Oxistop coatings will maintainoptimum heat transfer into boiler water wall tubes by replacing the iron oxide protective (andinsulating) layer and residue build-up that occurs when boiler tubes are subjected to high heatand coal combustion with a ceramic based protective barrier. This protective barrier will resistslag buildup that will also decrease heat transfer through the boiler tube.

History and use

Oxistop, LLC was formed in the spring of 2004 by a number of partners with extensive powergeneration maintenance problem solving experience. The company is the exclusive distributor andinstaller of Oxistop coatings and to date, over 30,000 Sq. Ft. of our coating has been installednationwide.


30/53



As a result of successful early trial applications and high interest in the industry in costeffective solutions to maintenance issues pertaining to boiler reliability, many repeat ordershave resulted from over sixty (60) initial applications in twenty-three (23) different states.

The ability of Oxistop high temperature metal coatings to reduce both slag and residue buildup,stop corrosion and oxidation and resist fly ash erosion and abrasion has led our client base toevaluate benefit regarding issues in the following areas:


31/53



BURNERS AND OVER FIRED AIRPORTS SLAG AND CORROSION.

REFRACTORY DOORS AND BURNER REFRACTORY SLAG

ECONOMIZER TUBES CORROSION AND EROSION

TUBE SHIELDS AND DUCT SUPPORTS CORROSION AND EROSION

NOSE AND SLOPE AREAS CORROSION AND SLAG

WATERWALL AREAS CORROSION, SLAG AND HEAT TRANSFER

SUPERHEAT AND REHEAT TUBES FOULING AND HEAT TRANSFER

To date, the types of coals and operations have had little affect on the success in coatingmaterial performance. Whether it is eastern or foreign coals, 100% Powder River Basin (PRB)or any blend combination, the materials perform equally well.This is not to say that Oxistop coatings cannot be tailor made to meet specific customerneeds.

Advantages of oxistopcoatings on metal

The advantages of Oxistop coatings compared to other coatings are a history of proven performance in the circulating fluidized bed boiler (CFB) industry for over fifteen (15) years.This material was enhanced and refined for specific applications in the mainstream pulverizedcoal market in October of 2004. Compared to metallic coatings such as thermal spray or weldoverlay, the advantages include not altering the integrity of the tube itself, thereby preventingor hindering future repair work to the tubes as the thermal spray and overlay process will.Stress is not put on the tubes as with an overlay process; in fact Oxistop coatings will improveheat distribution. Oxistop coatings can be removed by grit blasting or grinding if required.Whereas thermal spray and overlay will reduce heat transfer, Oxistop coatings will enhance

this process.

Oxistop Coatings Will:

Reduce the oxidation of metals at high temperatures.

Improve the temperature uniformity of boiler waterwall tubes.


32/53



Reduce the abrasive wear of fly ash on boiler tubes.

Reduce the buildup of combustion by-products in pulverized coal burning boilers.

Improve heat tr ansfer into boi ler waterwall tubes.

Demonstrate excellent corrosion and acid resistance at high temperatures.

Oxistop coatings are installed 3 to 6 Mils thick and will not crack due to expansion andcontraction of the metal substrate. They can be used in conjunction with thermal spray


33/53



technology, which will provide a cost effective package and act as a sealer for specific andsevere applications where the attributes of both, in combination, are an advantage.

Case studies

Although boiler runs between planned outages of 1236 months make evaluation of oxistopcoating trials time consuming, a number of repeat applications have been seen based onfavorable observations and measurable data. Examples are as follows:

In June of 2004, the first trial of Oxistop MC-19-GR coating was installed in a 100 MW boiler burning eastern bituminous coal. A waterwall test area of approximately 120 square feet and two

(2) burners of a sixteen-burner wall were selected to evaluate slag-shedding capabilities of thecoating. Row B burners were identified and coated as eyebrow slagging was heaviest a that level.Economizer tubes were also coated to eliminate corrosion occurrence between tube shields and thetubes.

Case study number one


34/53



After one year of boiler operation, the coated burners developed no eyebrows and were slagfree. During a planned outage in June of 2005 the coating material was found to be in-tack andcontinues to be in operation and performing into year three. As a result of this performance, theentire burner wall was coated (16 burners) and has continued to provide to date slag shedding

benefits, in addition to corrosion and erosion wear protection. Oxistop a GY-30 RefractoryCoating was utilized to provide a slag shedding protective barrier to the burner refractory. Alsocoated were additional waterwall areas, for slagging and corrosion, and the underside of theupper slope to eliminate clinker buildup.

Oxistop, Year Two:A Case Study

In April of 2006, Oxistop installed MC-19-GR coating on a 192 MW boiler burning 100%

PRB coal. Oxistop prepared by grit blast and coated less than half of the main combustion areaand most of the radiant zone of the reheat furnace from below the burners to above theoverfired air ports. This totaled a little over 5,500 square feet. The application was comfortablycompleted in the time allotted.

After approximately twelve days of base line information compiled in March, the preliminarydata verified that the application of Oxistop coating enhanced heat transfer in the coated arearesulting in an increase in boiler steam output of 5-7 MW. The FEGT (Furnace Exit GasTemperature) has decreased by 75 degrees. All this has been accomplished burning slightly

less coal. The waterwall tubes are cleaner and protected from fly ash erosion, corrosion, slag,and water cannon damage. Payback for the cost of the Oxistop coating application was 24days. Continual monitoring verifies heat transfer data indicating power output increaseremaining constant after more than five months on-line as a result of the Oxistop coatingapplication.

The Owner states the following:

Case study number two


35/53



1) The water cannon usage in this area is significantly lower.

2) The cordial thermocouple rate of delta-t is notably slower and, more significantly,

stabilizes at a lower number.

3) The furnace exit gas temperatures of the reheat box are below historical data.

4) High temp RH and low temp SH thermocouples are below historical data.

5) Decrease in air heater out let gas temperature.

6) Lower fuel usage eliminating over firing to make steaming rate.

7) Maximum megawatt output is increased to 198 Mega Watts, with slightly less coal burned.

A subsequent application of 5,520 square feet is scheduled for installation in the superheat boiler in the fall of 2006 based on impressive results of the spring trial. Similar benefits areanticipated.

Refractory coating applications

Utilized primarily around burners and refractory lined manway doors, Oxistop GY-30 is a3,000F coating material designed to give refractory substrates similar slag sheddingcharacteristics to the Oxistop metal coating.

Recently, Oxistop has developed a system using an unique refractory material that chemicallysets without heat and allows the Oxistop GY-30 coating to remain completely bonded afterinitial refractory curing. This is important during outages where time is of the essence and aone step package is required to address refractory repairs as well as the slag protection beingsought. As a result of packaging the coating with this impressive material, GY-30 sales have

accelerated.


36/53




37/53



Limitations

Oxistop MC-19-GR is a high temperature, corrosion resistant material that is used onwaterwall and related tube apparatus, such as superheat and reheat tubes where the range ofsteam temperatures are not limiting factors.

It had been used in conjunction with the CFB power industry, beginning in approximately1990, to minimize fly ash erosion, which approximates a lowgrade sandblasting affect. When the erosion resistance limits of the ceramic coating arereached, thermal spray technology has been used as a base and packaged with ceramic coatingto attain excellent results.

The non-catalytic nature of Oxistop coatings provides a protective barrier that prevents the

reaction of combustion by-product adherence to boiler tubes. Upper slope areas and wing-walltubes are some areas that have demonstrated the limits to Oxistop coatings in overcoming slagaccumulation due to boiler design issues. On low angle upper slopes, gravity and gas flow

patterns can overcome the advantages of this non-catalytic feature and allow for slag to remaina problem. Some wing wall designs that have small gaps between the tubes when surrounded

by combustion by-product gas flow streams, allow for slag to bond to itself between the coatedtubes to create the base for clinker formation. Coating of these tubes, in this case has shown tohave little effect on clinker formation.

Even with these observed limitations, slag resistance of Oxistop coatings has been readilydemonstrated around burners, over fired airports, on waterwalls, superheat and reheat platenand pendant tubes and the undersides of slope areas. Corrosion, tube wastage and erosioncontinue to be minimized and in many cases, eliminated by the coating. However, in that thesematerials rely on an initial heat curing for bonding and strengths, care must be given when the

practice of washing down scaffolding after an outage is used. After air-drying and before theheat set the coating will revert back to a wetted state if exposed to water. After initial heatcuring the coating becomes resistant to all acids, salts, solvents, water, and alkalies.

Soot blower type, velocity and particulate pickup during operation, all are mitigating factorsregarding eventual wearing of Oxistop coating around soot blowers. Although defining thelimits of ceramic coating are inconclusive and ongoing in regards to durability in these areas,thermal spray coating is a proven solution to this problem when combined with the Oxistopcoating.

Given the range and severity of maintenance issues regarding tube integrity and boilerreliabilityoverlay technology has, and will remain an option for restoring lost tube wall


38/53



thickness. Oxistop coatings have been introduced to the power industry to expand the range ofoptions for tube maintenance practices with additional benefits, including cost effectiveness.

Conclusion

At a cost of approximately one tenth of weld overlay, the industry interest continues to climbregarding both the Oxistop brand of metal and refractory coatings. While not marketed or seenas a cure-all, Oxistop coatings do provide proven alternatives and options to current practices

pertaining to high cost maintenance as well as boiler reliability issues. Oxistop coatings areuser friendly due to the fact that they are non-toxic, air drying and quickly applied. Curing ofthe material is accomplished during boiler start-up with no holds or separate heatingrequirements. Consequently, other repair work in the boiler may continue during theinstallation process. Expanded use of Oxistop coatings will reduce maintenance costs and

increase boiler efficiency and reliability over time. Durability has been demonstrated in initialOxistop coating trials, beginning in 2004. Each subsequent application has shown staying

power, in many cases outage-to-outage, and continued effective performance to date.

*ADDITIONAL PICTURES, PRODUCT INFORMATION AND USE CAN BE OBTAINED BY ACCESSING THEOXISTOP WEBSITE AT OXISTOP.COM. REFERENCE ARE AVAILABLE UPON REQUEST.

Tim E. BattonPresident and Senior Managing Partner

Al BraceyDirector of Materials and Installation

6331 Market StreetBoardman, Ohio USA 44512

This technical paper, and other Oxistop archives are available for download at www.oxistop.com.

The authors

Electronic access


39/53

Zhou Session 4 Straw Fired Boiler


Shedding of ash deposits in a straw fired boiler

H. Zhou, P. A. Jensen 1, F. Frandsen, J. Hansen, A. Zbogar

CHEC Research Center, Department of Chemical Engineering, Technical University of Denmark, Building 229, DK-2800 Kgs, Lyngby, Denmark

Abstract:The amount of ash deposit present on heat transfer surfaces is influenced by formation

and shedding processes. The shedding process was investigated in the Danish Avedrestraw fired boiler by probe measurements and development of a computer model describingdeposit formation and shedding from a superheater tube. Measurements with and withoutsootblowing were performed in the furnace where the flue gas temperature from 900 to1100 C and in the convective pass where the flue gas temperature from 650 to 830 Crespectively. The main shedding mechanism was observed to be melting on the deposit

surface, and subsequent ash droplet detachment in the furnace. The flue gas temperatureinfluences the melt fraction in the deposit and thereby, to a high degree, controls theshedding rate in the furnace. The mechanistic shedding model describes deposit formation

by condensation, ash particle impaction, and thermophoresis, and shedding by melting andsubsequent droplet detachment. The model was validated with the probe measurements andwas used to calculate the influence of changed local parameters on the build-up of thedeposit. Sootblowing was the main mechanism of the deposit removal in the convective

pass. Well-controlled sootblowing measurements were performed in order to determine thePeak Impact Pressure (PIP) needed to remove the deposits. The influence of a changed

probe metal surface temperature, and the residence time was investigated. At a probetemperature of 400 C, the deposit could easily be removed by the plant sootblowers. At a

probe temperature of 500C the removability of the deposit by sootblowing decreased withtime. The observed behavior of the deposit on the probe was generally similar to the boilerdeposits, at equal local conditions.

Keywords: Biomass combustion, Deposit, Shedding, Straw, Sootblower

1. IntroductionThermal conversion of biofuels to produce steam and power has emerged as a viable

alternative to the consumption of fossil fuels. Wheat straw is available in Denmark as one ofthe main biofuels and may be applied as a fuel in power plants. High concentrations of criticalelements such as potassium (K) and chlorine (Cl) in straw fuels have been established as themain reason for the serious problems with slagging, fouling, and corrosion encountered instraw fired boilers (Jensen et al. 1997, Baxter et al. 1998, Frandsen, 2005).

Several studies have been conducted on deposit formation and shedding in straw-fired power plant boilers. Jensen et al. (1997) reported that deposit fluxes are in the range of 15 to 160 and2 to 25 g/m 2h in the furnace where the flue gas temperature is around 850 oC and in theconvective pass where the flue gas temperature is around 650 oC of two Danish straw-fired

1 Corresponding author: Email [email protected] (PA Jensen), tel. +45 45 25 28 53


40/53



power stations. Deposit shedding experiments have been performed by Zbogar et al. (2006) inthe furnace of the straw fired (105 MW th) grate-boiler at the Danish Avedre power plant. Theexperiments indicated that the behavior of the deposit shedding on the superheaters mainlydepends on the local flue gas temperature. For instance, at flue gas temperatures as high as1100 oC or above, the deposit surface layer was completely molten and flowed around a

deposit probe, finally detaching the probe as well-defined droplets.Generally, sintered deposits are observed in the top of the second draught where the flue gastemperature is sufficiently low so that the deposit is not removed by surface melting. Depositsformed on the superheaters further upstream in the boiler is removed mainly by deposit surfacemelting, further downstream, the flue gas temperatures are usually so low that the deposit isoften regarded to be loose and easy to remove. Sootblowers are hence applied in the top of thesecond draught in order to assist the deposit removal in straw fired boilers. The objective ofthis work has been to investigate the behavior of deposit shedding in the furnace and theconvective pass of a Danish straw fired grate boiler. A combined air- and water-cooledshedding probe and support equipments such as a CCD camera and a sootblower probe wereused. The weight of the probe deposit and the probe heat uptake could be measured by the

shedding probe. Probe measurements were conducted to register the process of depositshedding, and to identify the jet Peak Impact Pressure (PIP) needed to remove the probedeposit as functions of the probe metal temperature and exposure time. Moreover, amathematical model has been developed for describing the deposit growth and shedding bydeposit melting and droplet detachment.

2. Experimental

2.1 The boilerA schematic diagram of the straw-fired grate-boiler, and the probe positions is shown in

Figure 1. The flue gas moves up through the furnace to the secondary superheater (SH2)

located at the top of the boiler, through the tertiary superheater (SH3) in a horizontal pass, andto the primary superheater (SH1) in the second pass. Fly ash particles rich in K, Si and Ca may be entrained from the grate combustion zone, and be transported with the flue gas, into the boiler chamber. The high alkali fly ash severely increases the deposit and corrosion problems,compared to boilers firing low-alkali biomass or coal.

The boiler was fitted with retractable sootblowers. The sootblowers extend into the boilerwhen active and retract outside the boiler wall when idle. The working fluid applied in thesootblowers is superheated steam at temperature of 573.15 K and a pressure of 19 bar. Therequired sootblowing frequency is 3 times per day. The plant sootblowing was activated onlyonce per day during the experiment.

2.2 The probe systemThe applied probe system is shown in Figure 2, and consists of the deposit probe, thesootblower probe, a CCD camera, an incandescent light source, and a separate thermocoupleused for measuring the flue gas temperature. A weight cell was connected to the probe so thatthe deposit weight could be registered on-line. The probe surface temperature was fixed by acombined air- and water-cooling system. The heat transfer to the probe was determined bymeasuring the air and water flows and their temperatures. Moreover, separate systems wereused for the extraction of fly ash particles and measuring flue gas velocity. More details can befound in Zbogar et al. (2006).


41/53



The sootblower probe nozzle is of convergent-divergent shape and the air at the nozzle exitis fully expanded at an air supply pressure of 6 bar. The removal of deposit by a sootblowermay be related to the jets Peak Impact Pressure (PIP), which is defined as the centerlinestagnation pressure of the jet (Jameel et al. 1994). Figure 3 presents the measured PIP, asfunctions of distance from the nozzle exit, and the supplying air pressure at room temperature

condition. The measured PIP increases with the supplying air pressure, and drops off quicklywith the axial distance from the nozzle exit.

2.3 The probe positions and local conditionsThe shedding probe positions in the boiler, are shown in Figure 1. The probe was situated

near the secondary pendant superheater in the furnace, and between SH1 and SH3 in theconvective pass. One of the plant sootblowers (see Figure 1) is located about 0.5 m above theshedding probe, in the convective pass when activated. Table 1 summarizes the localconditions at the shedding probe positions. The furnace has a higher flue gas temperature, ahigher flue gas velocity, and contains larger fly ash particles than those in the convective pass.The measured fly ash concentrations in the furnace and the convective pass are almost equal.

Table 2 summarizes the shedding measurement performed. Experiments were in 2003-2005.The total exposure time of the shedding probe in the boiler was about 2500 h. Probe metaltemperatures of 500 oC, and 400 to 550 C were applied in the furnace and in the convective

pass respectively. The probe exposure time at different probe metal temperatures varied from 3to 30 days. In two of the measuring campaigns (number 1 and 2) the boiler was stopped twoand three times respectively.

Simultaneously with the probe shedding experiments, the straw fired was sampled daily.Table 3 lists the proximate and the ultimate analysis data of the fuel. It contains 0.36 wt % Cl,0.92 wt % K and 1.1 wt % Si, which are typical levels of Danish wheat straw.

3. Results and discussion

3.1 Shedding of boiler tube depositsThe deposit shedding from the superheater tubes was registered by focusing the CCD

camera on the tubes. Figure 4 presents shedding of deposits from pending tubes in the furnace.The flue gas temperature is around 1200 oC and the steam temperature is 475 oC. The timerange of Figures 4 (a) and (b) is 12 minutes. It can be observed that a massive piece of moltendeposit rolled over the tube surface and moved from the upper to the low side against thedirection of the flue gas flow by its gravity, while some deposits remained on the tube surface.

Figure 5 shows the deposit accumulation on the superheater tubes in the convective pass.The flue gas temperature is about 800 oC and the steam temperature is 400 oC. Themeasurement was performed after the outer layer deposits were removed during a boiler shot

down. The time interval of Figures (a) and (b) is 7 days. Figure 5 (a) shows the formed depositon the superheater tubes 7 days after the boiler was restarted. It can be observed that anelliptical and dark grey deposit was formed at the upstream side of the tubes, while a thin greydeposit wrapped the downstream side of the tubes. The thickness of the deposit increasedgradually with time in the following days, and some bridging deposits were formed as shownin Figure 5 (b). It seemed like the plant sootblower could not keep a stable low amount ofdeposit if boiler shut downs, or manual deposit removal, was not applied.


42/53



3.2 Probe deposit shedding in the furnace

3.2.1 Experimental observationFigure 6 shows an example of an image of shedding from the deposit probe when the probe

was placed in the furnace. The droplets seen always detached at the bottom edge of the probe.

The shedding behavior depends on the flue gas temperature. It was observed that below a fluegas temperature of approximately 1100 oC, the deposit contained a certain solid fraction,whereas above 1100 oC, the video recording indicated that the deposit surface was completelymolten. At temperatures approximately 1100 oC, the melt flowed around the probe slowly,since solid particles increase the viscosity of the melt. At a temperature of approximately 900oC, molten phases and shedding by droplet detachment were rarely observed. This difference inmelt flow, can also be detected as a sudden decrease in mass loading on the probe, when thegas temperature increased to above 1100 oC. It was found that the change in deposit mass,caused by an increase in the flue gas temperature, is very fast, i.e. the response to the increasedflue gas temperature is almost immediate. Details are given in Zbogar et al. (2006).

3.2.2 Modeling workTo further investigate mechanisms of the deposit shedding in the furnace, a mechanistic

model including submodels of deposition (impaction, thermophoresis, Brownian and eddydiffusion, and condensation) and shedding by deposit surface melting has been developed. Thedeposit buildup rate at an angular position ( ) on the shedding probe is the difference betweenthe deposit growth and shedding rates, and can be expressed as:

( ) ( ) ( ) ( ) ( ) ( )

,,),(,,,,

t S t S t I t BE t TH t C dt t dm

f ++++= (1)

where m is the mass of deposit on the probe, t is the time, C, TH, BE, and I represent thedeposition rates by condensation (C), thermophoresis (TH), Brownian and eddy diffusions(BE), and impaction (I) including deposition of large particle inertia impaction on the upstreamside of the probe, and deposition of intermediated-sized particle by vortex interaction on thedownstream side of the probe. S is the shedding rate by drop detachment, and f S is the mass

accumulation rate in a control domain caused by melt film movement. More details of theshedding model are provided by Zhou et al. (2006).

Figure 7 compares the measured and the calculated probe deposit weight, and the heatuptake as a function of time, as well as the measured, and the applied input temperature.Generally, the predictions are in reasonable agreement with the measurements. Three distinct

periods can be distinguished in Figure 7, i.e. a fast increase of the deposit weight and acorrespondingly quick decrease of the heat uptake at an early stage (up to 10 h), then a slowincrease of the deposit weight and a decrease of the heat uptake up to 285 h, and finally amaturation period in which constant values of the deposit weight and the heat uptake appear. Asharp decrease of the deposit weight, and an increase of the heat uptake can be observed whena peak flue gas temperature at 1423 K applied. A discrepancy between the calculation and themeasurement at an early stage (up 80 h) is observed in Figure 7. This is mainly attributed tolarge uncertainties in some of the applied submodels.

The shedding model is capable of predicting not only the dynamics of the deposit weight andthe heat uptake as shown in Figure 7, but also the deposit shape, and the contributions todeposit formation by different mechanisms. A parameter study has been used to quantify theeffects of changed local conditions such as the flue gas temperature and velocity, the probe


43/53



metal temperature, and the entrained amount of straw ash on the probe deposit formation andshedding processes.

3.3 Removal of probe deposits in the convective passFigure 8 shows the variations of the measured flue gas temperature, the deposit mass, and

the heat uptake as a function of time in the period of 166 to 190 h, after the probe was initially placed in the boiler. The probe and the plant sootblowing were performed at 166.7 h and 167.3h respectively. The jets from the sootblowers caused strong fluctuations of the deposit massmeasurements. The flue gas temperature was quite stable before the boiler was shutdown at182 h. Some of the deposits were removed by the plant sootblowing, thereby the heat uptakeincreased slightly. The deposit mass increased noticeably between the plant sootblowing

periods. The measurements provide information about the change of the deposit mass and theheat uptake by probe during plant sootblowing, but it was not possible to quantify the amountof deposit removed by the individual probe sootblowing. The probe sootblowing events causeda drift in the baseline of the weight measurements, which made it impossible to determine theweight change accurately.

Figure 9 (a) shows the dynamic process of the removal of the probe deposit by the probesootblowing, two days after the deposit probe was placed in the boiler. The white pieces flyingaway from the deposit probe denotes the removed deposit, and the background of the figure isthe superheater tubes covered by deposit. The deposit surface became light grey as the cool air

jet impinged the deposit. The removal of the downstream deposit happened mainly as powders being blown away while the upstream deposit was removed in the large chunks, denoting thedeposit removal mechanisms by brittle fracture and debonding. Figure 9 (b) shows theconfiguration of the probe deposit after the probe was removed from the boiler. Along the

probe, the surface has a camelback-like shape. The formation of the shape may arise from theinfluence of the surrounding superheater bundles on the local flow field.

The outer layer of the downstream deposit on the probe was always removed easily, and a

portion of or no upstream deposit could be removed by sootblowing in the experiments. Theremovability of the deposit is therefore defined as the ratio (area, %) of the upstream outerlayer deposit that was removed by the sootblowing. The amount of the removed deposit wasroughly estimated from the CCD video images. Figure 10 shows results regarding depositremoval during the 740 hour long measuring campaign 5. The relative amount of probe depositthat could be removed by either plant sootblowing or probe sootblowing is shown. The top partof Figure 10 shows the percent of deposit removed by the plant sootblowing, and the bottom

part shows the level of probe jet PIP needed to remove the upstream deposit. For the plant sootblowing, less and less upstream deposit was removed with increasing

exposure time, before the first boiler shutdown (at 180 h). Almost all the upstream depositswere removed at t167 h. Boiler shutdown is beneficial to thedeposit removal, but the benefit is smaller at subsequent boiler shutdowns.The probe sootblower was during the first 350 hours operated at a distance of 20 cm from

the deposit probe, so a maximum PIP of 45 kPa was available. However, later the sootblower probe was moved closer to the deposit probe, and a maximum PIP of 190 kPa was obtained.For the probe sootblowing, the upstream deposit became harder and harder to remove beforethe first boiler stop (


44/53



Generally, it was observed that increased residence time or probe temperature meant thatless upstream outer layer probe deposit could be removed. For instances, at a probetemperature of 400 C, all deposit could be removed even at moderate PIP values. At 450 and500 C the PIP required increased with exposure time, and at 550 C even with the maximumavailable peak impact pressure of 190 kPa, no deposit could be removed.

ConclusionsThe deposit shedding in a straw-fired grate-boiler has been investigated experimentally and

by modeling. Both quantitative and qualitative information of the shedding processes wereobtained. The observed behavior of the deposit shedding on the probe generally was similar tothe boiler superheater tube deposits, at equal local conditions in either the furnace or theconvective pass.

In the furnace, where the flue gas temperature (900 -1100 oC) was high enough to causedeposits to be partially or fully molten, droplet detachment was the main mechanism ofdeposit shedding. The local flue gas temperature affects the deposit shedding rate. A higherflue gas temperature caused a high deposit shedding rate, a lower deposit mass, and a higher

heat uptake. The mechanistic model developed, including submodels of condensation, ash particle impaction, and thermophoresis, and shedding by droplet detachment is capable ofdescribing the observed deposit formation and shedding process. Furthermore, the model wasvalidated by the probe measurements and was used to calculate the influence of changed local

parameters on the development of the deposit weight and heat uptake.In the convective pass where the flue gas temperature (650 830 oC) is relatively low that

the deposits are mainly in the solid phase, the removal of the deposit is mainly caused byexternal forces i.e. sootblowing. The well controlled probe sootblower quantitativelydetermined the values of PIP needed to remove probe deposits as functions of the probe metaltemperature and the probe residence time. A lower probe metal temperature and a shorterexposure times leads to a lower PIP needed to remove the probe deposit.

AcknowledgementsThe deposit probe study was funded by PSO Energinet.dk. The company Dong Energi A/S

is acknowledged for letting us use the straw-fired boiler. This work is part of the CHEC(Combustion and Harmful Emission Control) research centre funded by the TechnicalUniversity of Denmark, Nordic Energy Research, Dong Energi A/S, PSO funds fromEnerginet.dk and the Danish Energy Research program.

References

1. Baxter, L.L., Miles, T.R., Jenkins, B.M., Milne, T., Dayton, D., 1998. The behavior ofinorganic material in biomass fired power boilers: field and laboratory experiences.Fuel Process. Technol. 54, 4778

2. Frandsen, F.J., 2005. Utilizing biomass and waste for power production a decade ofcontributing to the understanding, interpretation and analysis of deposits and corrosion

products. Fuel. 84 (10), 1277-12943. Jameel, M.I., Cormack, D.E., Tran, H.H., 1994. Sootblower optimization 1.

Fundamental hydrodynamics of a sootblower nozzle. Tappi Journal. 77 (5), 135-142

4. Jensen, P.A., Stenholm, M., Hald, P., 1997. Deposit investigation in straw-fired boiler.Energy & Fuels. 11, 1048-1055


45/53



5. Zhou, H., Jensen, P.A., Frandsen, F.J., 2006. Dynamic mechanistic model ofsuperheater deposit growth and shedding in a biomass fired grate boiler. Submitted toFUEL for publication.

6. Zbogar, A., Frandsen, F.J., Jensen, P.A., Glarborg, P., Hansen, J., 2006. Experimentalinvestigation of ash deposit shedding in a straw-fired boiler. Energy & Fuel. 20 (2),

512-519.


46/53



Figure CaptionsFigure 1. Schematic diagram of the grate fired boiler at the Avedre power plantFigure 2. The applied shedding probe systemFigure 3. Measured PIP as functions of distance from the nozzle exit and the supplying air

pressureFigure 4. Deposit shedding on pending superheater tubes in the furnace (campaign 1)Figure 5. Deposit formation on horizontal superheater tubes in the convective pass (campaign5)Figure 6. Probe deposit shedding in the furnace (campaign 1)Figure 7. Comparisons of the evolutions of calculated and measured probe deposit weight andheat uptake as a function of time in the furnaceFigure 8. Examples of the measured signals of the flue gas temperature, the deposit weight,and the heat uptake during experiments with a probe temperature of 500 oC in the convective

pass (campaign 5)Figure 9. Removal of deposit by artificial sootblowing and configuration of probe deposit after

the probe was removed from the convective pass (campaign 5)Figure 10. Relative amount of deposit by plant sootblowing and the level of probe jet PIPrequired to remove the front deposit (PIP values are the ones measured in laboratory at 20 oC,campaign 5)


47/53



Table 1: Measured local conditions in the furnace and convective pass

Furnace Convective pass

Flue gas temperature, oC 900 - 1100 650-830Flue gas velocity, m/s 10 4-7

Fly ash concentration, g/Nm 3 1.2 1Average particle size, m 58 35

Table 2: Performed shedding measurements

Number/Period Probe metaltemperature

(C)

Total hoursin the boiler

(h)

Boilershut-

downs

Number of plant

sootblows

Number of probe

sootblowsFurnace 1. 05/11-23/11/2003 500 376 2 -- --

2. 08/06 28/06 2004 400 A 454 0 12 0

3. 13/10 27/10 2005 400 B 332 0 12 54. 31/10 11/11 2005 450 263 0 10 85. 12/09 13/10 2005 500 A 743 3 22 166. 11/11 -14/11 2005 500 B 72 0 3 0

Convective pass

7. 14/11 27/11 2005 550 272 0 10 3

Table 3: Straw chemical composition (d.b., weight %)

Moisturecontent

Ash Volatile Fixed-C

10,2 4,7 76,6 18,7 C H N Al Ca Fe K Mg

47,23 6,34 0,406 0,0084 0,34 0,0077 0,92 0,065 Na P Si S Cl

0,024 0,060 1,1 0,093 0,358


48/53



Figure 1. Schematic diagram of the grate fired boiler at the Avedre power plant

SH2

SH3

SH1 probe infurnace

grate

straw

secondary air

primary air

probe inconvective pass

sootblower


49/53



1: balance weight, 2: load cell, 3: CCD camera, 4: boiler wall, 5: artificial sootblower probe, 6:

deposit probe, 7: TC for flue gas temperature measurement, 8: flange, 9: probe hinge

mounting, 10: incandescent light

(Note: 5 and 10 were not used during the shedding experiments in the furnace)

Figure 2. The applied shedding probe system

1

2

3

4

5

6

5, 6

78

9 10


50/53



0 100 200 300 400 500 600

0

50

100

150

200

250

300

350 5 cm 7 cm 9 cm

11 cm 13 cm 15 cm 20 cm

P I P p r e s s u r e

/ k P a

Supplying air pressure / kPa

Figure 3. Measured PIP as functions of distance from the nozzle exit and the supplying air

pressure

(a) falling deposit at 19:17 (b) falling deposit at 19:39

Figure 4. Deposit shedding on pending superheater tubes in the furnace (campaign 1)

deposit

flue gasflue gas

deposit


51/53



(a) - deposit shape on 23/Dec. (b) deposit shape on 30/Dec.

Figure 5. Deposit formation on horizontal superheater tubes in the convective pass

.

Figure 6. Probe deposit shedding in the furnace (campaign 1)

remains

sheddingdrops

flue gas flowdirection

flue gas flowdirection


52/53



Figure 7. Comparisons of the evolutions of calculated and measured probe deposit weight and

heat uptake as a function of time in the furnace

1

2

3

4

M a s s

/ k g / m

2

200

400

600800

F l u e g a s

t e m p e r a

t u r e

/ o C

boiler shutdown

164 168 172 176 180 184 188 1920

1020304050

Time / h

H e a t u p t a k e

/ k W

/ m 2

Figure 8. Examples of the measured signals of the flue gas temperature, the deposit weight, andthe heat uptake during experiments with a probe temperature of 500 oC in the convective pass

(campaign 5)


53/53


(a) removal of probe deposit (b) configuration of probe deposit

Figure 9. Removal of deposit by artificial sootblowing and configuration of probe deposit after

the probe was removed from the convective pass (campaign 5)

0 100 200 300 400 500 600 700

0

50

100

150

200

PIP at which the front deposit starts (0-25 %) to be removed PIP at which most of the deposit (>75 %) could be removed represents less than 25 % of deposit was removed at max. PIP

P I P / k P a

Time / h

0

20

40

60

80

100

R e m o v e d

d e p o

i s t / %

(a): little deposit left

by the artifical sootblowing

boiler shutdown

Figure 10. Relative amount of deposit removed by plant sootblowing and the level of probe jet

deposit probe

probe withsootblower nozzle

Flue gas

probe with deposit

upstream

downstream

Snowbird Jr m

Documents

Transcript of Snowbird Jr m