Scholarly article on topic 'Control system for an oxy-fuel combustion fluidized bed with flue gas recirculation'

Control system for an oxy-fuel combustion fluidized bed with flue gas recirculation Academic research paper on "Chemical engineering"

CC BY-NC-ND
0
0
Share paper
Academic journal
Energy Procedia
OECD Field of science
Keywords
{}

Abstract of research paper on Chemical engineering, author of scientific article — I. Guedea, I. Bolea, C. Lupiáñez, N. Cortés, E. Teruel, et al.

Abstract In oxy-fuel combustion it is generally accepted that fed O2 into the boiler diluted together with CO2 is advantageous when similar combustion conditions as in air-firing case want to be reached. In experimental oxy-fuel combustion plants, auxiliary CO2 from commercial tanks is often used for carrying out experimental tests. However, this fact entails high operational costs in addition to scale prices are not applied in experimental size plants. A reduction of these costs could be achieved by supplying the required CO2 through flue gas recirculation, as it would happen in an industrial oxy-fuel plant. However, transition from air-firing to oxy-firing, and changes in the flue gas composition during operation can vary conditions inside the boiler. This issue becomes particularly important when oxy-fuel combustion takes place inside a fluidized bed boiler, where fluidization conditions must be kept inside a narrow optimum range, i.e. fluidization velocity or oxygen content at inlet, although changes in recirculation rates or flue gas composition could disturb it. With this aim, an efficient recycled flue gas control strategy has been modeled and validated through experiments in a test rig. This system could be easily adapted to any experimental oxy-fuel pilot plant, since it has been designed for controlling conventional instrumentation, valves and drives.

Academic research paper on topic "Control system for an oxy-fuel combustion fluidized bed with flue gas recirculation"

Available online at www.sciencedirect.com -

% ScienceDirect Ener9y

Procedía

Energy Procedía 4 (2011) 972-979 :

www.elsevier.com/locate/procedia

GHGT-10

Control system for an oxy-fuel combustion fluidized bed with

flue gas recirculation

I. Guedea, I. Bolea, C. Lupiáñez, N. Cortés, E. Teruel, J. Pallarés, LI. Díez, LM. Romeo

Center of Research on Energy Resources & Consumption (CIRCE). Universidad de Zaragoza. Mariano Esquillor, 15, 50018 Zaragoza (Spain)

Abstract

In oxy-fuel combustion it is generally accepted that fed O2 into the boiler diluted together with CO2 is advantageous when similar combustion conditions as in air-firing case want to be reached. In experimental oxy-fuel combustion plants, auxiliary CO2 from commercial tanks is often used for carrying out experimental tests. However, this fact entails high operational costs in addition to scale prices are not applied in experimental size plants. A reduction of these costs could be achieved by supplying the required CO2 through flue gas recirculation, as it would happen in an industrial oxy-fuel plant. However, transition from air-firing to oxy-firing, and changes in the flue gas composition during operation can vary conditions inside the boiler. This issue becomes particularly important when oxy-fuel combustion takes place inside a fluidized bed boiler, where fluidization conditions must be kept inside a narrow optimum range, i.e. fluidization velocity or oxygen content at inlet, although changes in recirculation rates or flue gas composition could disturb it. With this aim, an efficient recycled flue gas control strategy has been modeled and validated through experiments in a test rig. This system could be easily adapted to any experimental oxy-fuel pilot plant, since it has been designed for controlling conventional instrumentation, valves and drives. ©2011 Published by Els evier Ltd.

Keywords: oxy-fuel combustion; fluidized bed; flue gas recirculation; control system.

1. Introduction

Oxy-fuel combustion is one of the most promising technologies enabling the economic and efficient carbon dioxide capture. For this reason, many leading companies and research groups are working in this line to increase its knowledge and experience. Oxy-fuel combustion is usually carried out by using a gas mixture of pure oxygen and carbon dioxide as comburent. Generally, the CO2 in the mixture comes from recycled flue gases (RFG) that have been produced in the boiler. This way, oxygen is diluted by a fraction of flue gas that returns to the combustion chamber, at the same time that an enrichment of carbon dioxide in the flue gas is achieved. Also other species concentrations, such as water vapour, and sulphur dioxide are affected by this process increasing their concentrations in flue gases. The other fraction of flue gases are conveyed then through a condenser, where water vapour is removed and the remaining stream consists mostly of CO2. Recent research studies have focused on combustion efficiency under different O2/CO2 atmospheres [1, 2], finding optimum recycled ratios [3, 4] or adapting technology designs to the comburent composition [5], but as far as authors know, there are not studies about how recirculation can be carried out

doi:10.1016/j.egypro.2011.01.144

and controlled. In this study, attention has been paid to a major aspect in the oxy-fuel combustion process: the control of the flue gases recirculation, which has an outstanding influence for reaching high combustion efficiency, and consequently, an adequate heat transfer rate in the boiler.

Flue gas recirculation control turns up to be even more important when using fluidized bed technologies, because the velocity inside the boiler, and hence, the comburent volumetric flow rate, is a critical operation parameter to obtain advisable fluidization conditions. Moreover, higher O2 concentration than with air combustion can be introduced into the boiler, so control over the recirculation ratio becomes an important issue, to maintain adequate conditions in the reactor for combustion and pollutant emission regulation. The main objective of this work is setting up an efficient control system of the recycled flue gas in experimental oxy-fuel combustion plants based on fluidized bed technologies. This will allow establishing desirable operation conditions and comparing them with experimental tests. To achieve this aim, firstly a model has been developed and several simulations with different O2/CO2 ratios have been carried out to estimate the needed recycled flow rate for the oxy-fuel combustion in a fluidized bed so as to predict the gaseous species that would be recycled. Experimental investigations have been carried out to verify the correct performance of the control system.

2. Experimental Facility

CIRCE Oxy-fuel Fluidized Bed test rig (CIRCE-OFB) consists of a 90 kW bubbling fluidized bed combustor of 203 mm diameter and 2.5 m high. The feeding system includes two hoppers with their respective screws, which facilitate the operation with fuel blends at the same time, such as coal and biomass. The oxidant stream is provided by a forced fan, when air-tests are carried out, and by canisters, when O2/CO2 gas mixture is used as comburent. The gas mixture is prepared in a lunge connected to the canisters, where O2/CO2 proportion can be fixed. An induced fan helps the flue gases to exit the system in both cases, after passing through a high efficiency cyclone, a heat recovery exchanger, and a bag filter. When recirculation is applied, the forced fan introduces part of the flue gases again in the boiler at the required proportion, by means of a frequency speed driver and automatic valves. Further details are explained in section 4.

3. Simulation results

Before designing the control strategy it is necessary to bear in mind how recycled flue gases affect the plant performance. The main parameter is the amount of oxygen required for the combustion, which is set by the fuel thermal input and the imposed excess of oxygen, over the stoichiometric. In the experimental plant, when no recirculation takes place, CO2 from cylinders must be mixed with the O2. It is generally accepted that a proportion of 30/70 of O2/CO2 in the inlet stream is needed under oxy-fuel combustion to reach similar conditions to those of air-firing [6]. However, when part of the flue gas is recycled, other gas species besides CO2 are present in the comburent stream. Significant amount of O2 will be contained in the RFG (around 3-6%). This O2 must be considered for decreasing the O2 needed from the cylinders. N2 will also be present in the flue gas stream. It comes from three main sources: O2/CO2 cylinders impurities, fuel nitrogen content and air leakages into the system. In the literature [7] a ratio of 3% is proposed as a minimum value under which avoiding leakage is not economically feasible. New plants could achieve this goal, but this value for retrofit plants is estimated between 8-16% [8].

Finally, it is important to consider steam accumulation in the system when flue gas is not dried before being recycled, as in CIRCE-OFB. Steam in the oxidant stream could affect the combustion efficiency, heat transfer and the sulphur capture i.e. the boiler performance. Some authors consider wet recirculation a promising alternative [9, 10] since preheating energy consumption could be reduced, but corrosion caused by condensation represents a drawback [7].

Simulations have been carried out for different fuel thermal inputs, O2 concentrations in the inlet and recycled ratios, before implementing the control strategy for the recirculation. The necessary O2 is determined by the fuel input for a proper combustion (stoichiometric oxygen and an oxygen excess to assure complete fuel combustion). Fluidization conditions in the boiler determine the volumetric flow rate of gases entering the boiler and hence, the amount of gas that should be introduced together with the required O2. To have a wide range of characterized situations, two different coals have been considered (lignite and anthracite), as well as recirculation ratios from 0 to 60% and O2 concentrations at the entrance of 21%, 30% and 40%. Flue gases in CIRCE-OFB test rig are not dried before recycling and steam content in flue gases increases with higher recirculation rates, as shown in Figure 1. The O2 and CO2 purity in the cylinders is 99.9%. The lambda is 1.15 (O2_inlet/O2_ stoichiometric) and, for this first study, minor components are considered negligible.

Some of the species concentration results are shown below (Figure 1) for a thermal input of 90 kW, and lignite as fuel. As it was expected H2O concentration increases and CO2 concentration decreases while recirculation rate is increased. Also O2 concentration increases while the recirculation rate does and O2 proportion in the oxidant stream is higher. That is due to the decrease of overall volume rate of flue gases, since less inert gas is introduced with the oxygen. Nitrogen change is not appreciable because by the moment air leakage is not considered.

Figure 1. Recirculation influence on outlet gas species for different O2 proportion.

The information taken from the simulation gives a range of data for establishing set-points of flow rates for the fluidized bed operation, which will be used by the controller to regulate gas flow rates coming from recirculation and from the cylinders, so that the flow rate fulfils the desirable conditions. For better understanding how the controller should operate, Figure 2 shows how the O2 and CO2 concentrations vary with flow rates once the recirculation takes place. In this example, an oxy-fuel combustion test 30O2/70CO2 is performed. At the beginning, the comburent stream is coming from the cylinders until the moment that 30% of the flue gases starts to being recycled. C1 is the gas flow at the boiler inlet; C3 is the recycled flow rate and Cbot is the flow coming from the cylinders.

Figure 2. Simulation of changes in O2 and CO2 concentration in C1, C3 and Cbot for 30% flue gas recycled.

In Figure 2 it can be seen that an increase in O2/CO2 ratio from cylinders is needed, for keeping constant the overall proportion of O2 entering the fluidized bed. That is also important in terms of reducing CO2 flow rates from cylinders, which means high operational costs for an experimental pilot plant. From these results the controller is programmed so that the fan drive speed, concentration in the lunge mixer and automatic recirculation valves can be modified and adapted during operation to satisfy the aforementioned requirements.

Simulations considering air leakage into the system have been also carried out since leakages have been detected in CIRCE-OFB previous oxy-fuel tests. In the graph shown in Figure 3 it can be seen how CO2 purity decreases when increasing the recirculation ratio. This is caused by an increase of the accumulation of impurities coming from air, basically N2.

100 90 "5 80 £ 70

£ 50 J3 40 .E 30 O 20 10 0

0 5 10 15 20

Leakage (%vol.)

■■No recirculation -♦■30% recycled 60% recycled

Figure 3. Air leakage influence on CO2 concentration in flue gases. 4. Control system design

Once the recycled flow is characterized by the above explained model, a flexible recycled system is needed to perform different operation conditions. Being able to use a wide range of coal and biomass feeding, oxygen inlet concentration from 21% to 40%, and having appropriate fluidization velocity during the tests are the challenges. Moreover, the control system should be able to change flue gas recycled ratios and, especially important, it should make a smooth transition from not recycle to recycle mode. When flue gas recirculation starts, the gas flow at the entrance of the bed, C1, must remain constant to ensure proper fluidization velocity. In order to maintain the reference value or set point, SP_C1, and, at the same time, reaching a quick adjustment, it is necessary to design and implement an automatic recirculation control. The proposed solution is based on conventional instrumentation and control tools, which can be readily adapted to other oxy-firing test rigs.

The recycled stream control is implemented in a medium-scale Programmable Logic Controller (PLC) that is in charge of controlling the whole facility, supervised by a Supervisory Control and Data Acquisition Application (SCADA). This control equipment, with capability to apply up-to-date regulation, automation, and monitoring methods, ensures the required performance and flexibility for implementing diverse control strategies. For the moment, there is a lack of information on how to develop an accurate control to RFG in pilot plants.

Instrumentation

Variable frequency drives are in charge of controlling fuel feeding screw motor speeds. Together with fuel type and composition, this provides the required thermal power for the combustor. From this point, O2 inlet concentration in the O2/CO2 mixed stream from the cylinders is chosen. CO2 concentration and pressure are measured and registered. Without recycled stream, these measurements together with the flow rate are enoueh to control the oxidant stream.

When recirculation takes places, gas flow from cylinders and recycled gas stream have to satisfy together the inlet stream requirements, for which fans and valves must be manipulated.

Flue gas specie concentrations are measured continuously and must be considered. These compounds are O2, CO2, CO, SO2 and NOx and they are measured in the combustor feeding pipe (recycled gases together with cylinders stream) and in the exit flue gas pipe. This is performed by an automated continuous emission monitoring system, with alternate sampling from two points. The next figure, Figure 4, shows a diagram of the main devices involved in recirculation:

Figure 4. Scheme of the main devices involved in recirculation control. System performance

The system is operated as follows: thermal power and inlet O2 concentration determine comburent flow. When recirculation takes place, the flow from the cylinders should be decreased and the O2 concentration should be increased in the lunge according to the imposed needs by the operating conditions. Then valve V10 is opened, valve V11 is partially closed and the RFG flow, C3, is adjusted by varying the frequency Ve_2 fan, so that flow C1 is keep at the reference value. The forced fan is started to achieve the SP_C1 and ensures this value constant against possible fluctuations. These fluctuations are pressure drop changes along the gas circuit, mainly across the bed and O2 concentration in exhaust gases. That is, the sum of O2 in the exhaust gases and O2 bottles should be the percent desired in the inlet stream.

Once the test rig operation and the influential variables in the process were analyzed, cold tests were carried out to study the dependence of recycled flow C3 with different variables of the plant as valves opening, V2, V10, V11 and fan speeds, Ve_2 and Ve_3. As it was expected the main parameter that disturbs C3 flow rate is the forced fan speed.

Regulation

The major goal of the control strategy is to provide an incoming comburent flow to the boiler with the desired O2 concentration, by adjusting the recirculation and the supply of oxygen. The control system manipulates:

- The speed of the draft and induced fans, by means of variable speed drives.

- Proportional control valves located in the exhaust and recycled gases pipes.

- The inlet pressure of the O2/CO2 mixture injected in the gas feeding pipe. The mixture is provided by the set of canisters, where the O2/CO2 ratio can be set.

After analysing the different influences, the forced fan speed, Ve_2, was selected as the most convenient variable to be manipulated for recirculation regulation, aiming at keeping the measured flow, C1(t), at its set-point, which is calculated under all requirements exposed above. As it can be seen in the scheme of Figure 5, a simple feedback loop with proportional-integral (PI) action is used. Derivative action to improve transient response is neither necessary, due to the high response velocity, nor convenient, due to fluctuations in the controlled variable. The diverse changes in the operation, such as fuel energy flow, inlet O2 concentration, recycled flow ratio and pressure in the cylinders flow pipe, can all of them be regarded as perturbations that the automatic regulation of the forced fan speed must compensate for.

[m3/h] C1

Sensor/Transducer

Figure 5. Block diagram of the fan speed regulation. 5. Recirculation tests results

For confirming the proper performance of the control strategy, tests were made with cold flow, i.e., without combustion. Results from these tests, consisting on stepping from 0% to 35% or 45% recycled flow rate, case 1 and case 2, respectively, are represented in Figure 6. In these graphics it can be seen how the system reacts to a step change in the valve that reduces cylinders flow (V2, see the scheme of Figure 4). Reacting to the (almost) immediate drop of the flow from bottles, the fan speed rises, increasing the recycled flow, with a response time about half minute. All in all, the flow at the boiler inlet is barely perturbed, the desired value is restored in that half minute, quick enough for having stable conditions inside the combustor, keeping the fluidization conditions almost unchanged when recirculation begins.

> 10 >

3500 3000 ¡^ 2500 2000

^ 1500 >

i 1000

O 20 _Q

70 60 J* 50 40 ~ 30 U 20 10 0

Time(s)

Time (s)

Figure 6. Step response of the recirculation control system.

With the aim of validating the control strategy during oxy-fuel combustion, oxy-firing test was carried out in the CIRCE-OFB test rig introducing RFG. The fuel used was lignite, a 30% O2 concentration was kept in the oxidant stream and a 30% of the flue gas was recycled. As explained above, a fast response of the forced fan speed regulation is desired to keep fluidizing velocity constant, and the necessary O2 rate for the combustion. In Figure 7 it can be seen that the set-point for the flow rate is recovered in 45 seconds after suddenly dropping due to the (partial) closing of the bottles valve that initiates recirculation. The bed temperature and pressure drop along the bed are also represented, together with the boiler inlet flow rate and recirculation fan drive speed. Smooth change of conditions at the boiler entrance is reached, keeping bed temperature and pressure drop unchanged, while RFG is mixed with the gas from the cylinders.

Flow rate (m3/s) Pressure drop (mbar) 45 40 35 30 25 20 15 10 5 0

Bed Temperature (0C) Fan speed (rpm) 4000

3500 3000 2500 2000 1500 1000 500 0

400 425 450 475 500 525 550 575 600 625 650 675 700 Time (s)

Pressure Drop Flow k Ve2 Speed —Bed temperature

Figure 7. Control performance in oxy-fuel test with RFG.

As it has been mentioned, in the test rig air leakage becomes a real problem when the flue gases are recycled since the CO2 concentration outlet decreases. A depressed pressure zone is present before the induced fan. Since N2 is not measured, the air leakage in the system during oxy-firing conditions was calculated. The result was an air leakage around 7-8%. From the oxy-fuel test, purity of CO2 decreased around 6 points at the flue gas outlet, when air leakage was 8%. This is remarkable for assessing the influence of recirculation in the flue gas species content, since CO2 purity is one of the challenges in oxy-fuel combustion for CO2 capture.

6. Conclusions

An efficient RFG control system has been designed and its implementation has been presented along this paper. A mass model has been the first step to develop and characterize the recirculation in CIRCE-OFB test rig with the main objective of setting up the main parameters of the recirculation system. The distributed instrumentation and control devices along the plant facilitate the oxy-bubbling fluidized bed reactor operation and control.

Required parameters for a smooth change to recycling-mode, from the non-recycled mode have been identified. These were the amount of needed oxygen for combustion and the gas flow rate entering the combustor for a proper fluidization. A recirculation control strategy has been successfully developed and simulated. Tests carried out in the CIRCE-OFB test rig confirmed the good performance of the controller. In spite of being designed for a specific experimental facility, the proposed control system and described strategy here can be easily adapted to any experimental plant that aims to apply flue gas recirculation in a fluidized bed. Besides the advantages of the RFG in the case of oxy-fuel combustion, a necessity of cost reduction in research plants makes it highlighting interesting due to its intrinsic reduction in the consumption of gases from commercial cylinders.

During recycled-mode, non-condensable gas concentrations are increased in the facility. The designed controller takes into account the oxygen from the recycled stream to reduce the consumption of O2 coming

from the cylinders and the accumulation of this specie during the tests. Non-condensable gases may increase compression work of the CO2 capture so air leakage and the oxygen excess in combustion flue gases must be kept in low levels and also the O2 purity introduced from an air separation unit or cylinders will introduce N2 or Ar into the system.

Acknowledgements

The work described in this paper was supported by the R+D Spanish National Program from Ministerio de Ciencia e Innovación, MICINN (Spanish Ministry of Science) under projects ENE2009-08246 and CIT-440000-2009-26. MICINN is gratefully acknowledged for the individual fellowship granted to C.L. Fundación CIUDEN is also acknowledged for the support to the oxyfuel test rig.

References

[1] Toporov, D, P Bocian, P Heil, A Kellermann, H Stadler, S Tschunko, M Förster, R Kneer, Detailed investigation of a pulverized fuel swirl flame in CO2/O2 atmosphere. Combustion and Flame, 2008. 155(4): 605-618.

[2] Borrego, AG, D Alvarez, Comparison of Chars Obtained under Oxy-Fuel and Conventional Pulverized Coal Combustion Atmospheres. Energy & Fuels, 2007. 21(6): 3171-3179.

[3] Croiset, E, KV Thambimuthu, NOx and SO2 emissions from O2/CO2 recycle coal combustion. Fuel, 2001. 80(14): 2117-2121.

[4] Buhre, BJP, LK Elliott, CD Sheng, RP Gupta, TF Wall, Oxy-fuel combustion technology for coal-fired power generation. Progress in Energy and Combustion Science, 2005. 31(4): 283-307.

[5] Scheffknecht, G. Research Perspective- Review of the Current Understanding, Identifying Research Gaps. in 1st Oxyfuel Combustion Conference. 2009. Cottbus, Germany.

[6] Santos, S. Challenges in the Development of Oxy-Coal Combustion Power Plant with CO2 Capture. in Asia Pacific Programme (APP). Oxyfuel Working Group Capacity Bldg. Course. 2009. Daejeon, South Korea.

[7] Toftegaard, MB, J Brix, PA Jensen, P Glarborg, AD Jensen, Oxy-fuel combustion of solid fuels. Progress in Energy and Combustion Science, 2010. In Press, Corrected Proof.

[8] Dillon, DJ, RS Panesar, RA Wall, RJ Allam, V White, J Gibbins, MR Haines, ES Rubin, DW Keith, CF Gilboy, M Wilson, T Morris, J Gale, K Thambimuthu, Oxy-combustion processes for CO2 capture from advanced supercritical PF and NGCC power plant, in Greenhouse Gas Control Technologies 7. 2005, Elsevier Science Ltd: Oxford. p. 211-220.

[9] Dillon, DJ, V White, RJ Allam, RA Wall, J Gibbins, Oxy-Combusion Processes for CO2 Capture from Power Plant. 2005, IEA Greenhouse Gas Research and Development Programme.

[10] Nakayama, S, Y Noguchi, T Kiga, S Miyamae, U Maeda, M Kawai, T Tanaka, K Koyata, H Makino, Pulverized coal combustion in O2/CO2 mixtures on a power plant for CO2 recovery. Energy Conversion and Management. 1992. 33(5-8): 379-386.