Feature: A Boiler Combustion Control System with Combustion Ratio Optimisation and Soft Measurements

A Boiler of control with Foundation Fieldbus in aspect of system

architecture and soft configuration are covered in Section
3. Control effects are described in Section 4.Conclusions

Combustion and recommendations are given in Section 5.

2. Measurement and Control Principles

Control System for a Boiler Combustion system
with Combustion Figure 1 shows the process for combustion system of
boiler. Steam is produced by a drum, which is heated by

Ratio Optimisation means of water wall around the inside of furnace. As for
the control schemes, three control loops are installed for

and Soft
the fuel, the airflow and the furnace negative pressure, and
the two inputs are the coal and air. The objectives of the
combustion control system are:

Measurements 1. to stabilize stream pressure

2. to achieve optimum combustion efficiency by automatically
searching and maintaining optimum air-to-fuel ratio Ro
3. to stabilize negative pressure in furnace.
Geng Liang, Yan Bai 1. Introduction
and Guotian Yang
School of Control and The level of automation in boilers is relatively low
Computer Engineering, and conventional control systems based on analogue
North China Electric Power instruments with dashboards are still widely used today.
University, China Investigation has shown that heat supply boilers without
any automatic measurement and control account for
Wen Li nearly 50% of the total number; and that by means of
Beijing GuoDianZhiShen Control conventional dashboard-based solution accounts for
Technology Co. Ltd., China 75-80% around the world. Some serious problems in the
control of medium and small capacity boilers are: Figure 1: Process for Boiler Combustion System
1. Mostly, combustion in boilers is manually controlled.
2. Manual operations in the running of combustion Three loops serves as fuel feeding control, forced
system are notoriously complex, often causing systems draft control and induced draft control respectively.
breakdown. Conventional cascade PID control strategy is used in the
3. Some semi-automatic control systems based on fuel control loop, with primary PID regulator stabilizing
analogue instruments and devices often lead to steam pressure and secondary PID regulator regulating
degradation in control effect and low reliability. coal feeding. Output of primary PID regulator is tracked
4. Self-diagnosis and test information from fields are by the secondary one. Steam flow and drum pressure are
unavailable with analogue communication applied in used to evaluate the heat release in process of combustion.
field, making the system monitoring and maintenance a Air-to-fuel ratio R, multiplied by the output of fuel
heavy workload. loop PID controller to serve as the setting point of the
5. Data transmission based on point-to-point topology in PID controller of air flow loop, is introduced to connect
field within control systems makes wiring and installa- the two loops so as to ensure efficient combustion. Air
tion of systems a costly, time-and-labour consuming task. pressure in inlet and outlet of pre-heater are used to
The advent and development of fieldbus technology pro- evaluate the airflow. The air-to-fuel ratio R is evaluated
vide ideal solutions to the renovation of factory automation and optimized with the algorithms proposed in this paper.
with its field-based digital communication, distributed con- Furnace negative pressure is regulated to track setting point
trol and information acquisition. In this paper, we present with induced draft regulator. Especially, a feed forward from
a comprehensive design and implementation of a system forced draft is adopted to stabilize furnace negative pressure.
for combustion control based on Foundation Fieldbus (FF). The complete control strategies are shown in Figure 2.
Designs of algorithms, implementations in terms of system
architecture and configurations are included. We propose Air-to-fuel Ratio Optimisation
a scheme for optimising air-to-fuel ratio and implement a Air is decisive in combustion efficiency. Generally, air is
number of function blocks in field intelligent devices. We required to change with fuel to achieve desired combus-
also propose an approach on soft measurement of airflow tion efficiency, which can be represented by heat release
based on the orifice nature of air pre-heater. loss in combustion.
This paper is organized as follows. The measurement Traditionally, the air-to-fuel ratio was set as a constant
and control principles for the combustion system and the in a combustion control system, which lacks adaptability
design and analysis of the proposed soft measurement and makes it difficult to keep an optimum ratio because
of airflow are presented in Section 2. Implementation of changes in combustion condition, often causing low

112 • Measurement + Control Vol 43/4 May 2010 www.instmc.org.uk

 Feature: A Boiler Combustion Control System with Combustion Ratio Optimisation and Soft Measurements

Figure 2: Control Strategies for the Designed System

combustion efficiency in practice. In this paper, two optimisation

algorithms for air-to-fuel ratio are proposed and used to search the
optimum air-to-fuel ratio in real time. Ratio control is then used to
adjust the airflow by regulating the forced draft. The optimum ratio
can be searched automatically with the two algorithms to keep
the combustion system operating either closely around or at the Figure 4: Flow Chart for Gradient Searching Optimisation Algorithm
optimum efficiency point.
A typical air flow-heat release curve is more often used in these methodologies, which make them difficult to be applied with
algorithm design, as shown in Figure 3, where Q stands for the heat field-based distributed control in practice. Inaccuracy in obtained
release, F represents the airflow. Air-to-fuel ratio R was a function model also causes degradation in measurement and control. As far
of airflow since fuel is regulated according to steam pressure and as the stated combustion system is concerned, regulation of airflow
can be taken as a constant at any given time instant. The optimum can necessarily lead to change of oxygen content although the latter
Air-to-fuel ratio Ro can be achieved when heat release reaches the is essential in combustion efficiency. Therefore, regulation of airflow
maximum with change of airflow. can be viewed as a collateral and effective means of control and
Maximum allowable progressive there is no need to measure oxygen content with the methodologies
air-to-fuel ratio increment ∆R is mentioned above since combustion efficiency is regulated by the
taken as the step-size in optimisa- airflow with the optimisation of air-to-fuel ratio. Only the airflow
tion. Gradient method is used needs to be measured as required by the optimisation.
to search the optimal ratio, with The most
airflow as controlling parameter common way to
and the heat release as controlled measure airflow
parameter.The optimal air-to-fuel is to use orifice
ratio Ro can then be obtained Figure 3: Air flow-heat Release flow meter albeit
when the peak point is attained Relationship Curve this can inevitably
in the curve in Figure 3. cause throttling loss.
The optimal point is on the right side of the current calculating Experiments show
point if the gradient was positive. Then increasing airflow will that throttling Figure 5: Structure of Tubular Air Pre-heater
cause the heat release to increase. Current air-to-fuel ratio will be effect takes place
increased next; the optimal point is on the left of current calculating when air flowing through the air pre-heater of boiler. The common
point if the gradient is negative. The airflow will be decreased next feature in various pre-heaters of boiler including cast iron, tubular
instead. The optimal R is searched with gradient algorithm. That is, the and heat-pipe ones widely used in medium and small capacity boilers,
air-to-fuel ratio increment ∆R is used in search a combining judge- is the heat exchange components. Heat exchange components can
ment of the changing trends of heat release and airflow. Flow chart be viewed as some type of cascade or paralleled orifices when air
for the optimisation is shown in Figure 4. flows through them as shown in Figure 5.
The advantages of this algorithm lie in its easy implementation and The general characteristic of differential pressure-air flow for an
less computation involved. air pre-heater is similar to that of the genuine orifice in trending.
Therefore, the air pre-heater can be viewed as a kind of specific
Measurement of Air orifice flow meter. Air pre-heater is natural and ideal for measuring
Oxygen content, which can be regulated by controlling airflow, is a the airflow going through it. Consequently, the soft measurement
vital parameter for the combustion efficiency. For the measurement method would help to reduce cost and avoid throttling loss caused
of oxygen content, most of methodologies are hardware-based, by additionally installed orifice flow meters. Airflow can be measured
such as Zirconia oxygen analyzer and some other advanced devices by means of measuring the air pressure difference before and after
[1]
. Instead of hardware-based measurement, more researchers the air pre-heater, i.e. inlet and outlet of air pre-heater, with differen-
are focusing on soft measurement of oxygen-content, where tial pressure detecting instruments.
neural network, SVM, statistical analysis and data fusion, dynamic
identification and modelling are the main methodologies used in the Measurement of Fuel
latest research.[2]. Some researchers proposed airflow modelling by Changes in fuel can lead to variations in the heat release of boiler
Computational Fluid Dynamics [3]. Nevertheless, Complex algorithms eventually. Heat release is the ultimate reflection of fuel. Therefore,
and massive computations involved are the common problems for heat release of boiler can be taken as a primary feedback represent-

www.instmc.org.uk Measurement + Control Vol 43/4 May 2010 • 113

Feature: A Boiler Combustion Control System with Combustion Ratio Optimisation and Soft Measurements

ing the fuel. The advantages of measuring fuel by means of heat boiler system. Industrial control computers including operator
release are as follows. workstation and engineer workstation, control scheme configuration
• Any change or disturbance in fuel transmission system has no software and HMI software installed, were used and deployed in
effect on accurate measurement of the fuel; supervisory central room. Supervisory computers and Ethernet/
• Suitable for measuring various kinds of fuel or mixed fuels and Fieldbus gateway DFI302 are connected by 10MB Ethernet. OPC
measurement accuracy can be guaranteed; Server was installed in supervisory computers to realize the com-
• Changes in heat-release of fuel have no effect on accurate munication between supervisory computers and gateway. Fuel supply
measurement of the fuel; was regulated by variable-voltage & variable- frequency (VVVF)
• Load varying has no effect on accurate measurement of the fuel. motor in the main steam pressure control system; forced draft was
Consequently, heat release can reflect the changes in the fuel into controlled by VVVF motor driving forced draft fans in forced draft
the furnace quickly and accurately. It does not need any correction control system; furnace negative pressure was controlled by VVVF
for the measurement caused by disturbances such as changes in fuel motor driving induced draft fans in furnace negative pressure system.
transmission, fuel types and heating value, etc. A typical FF-FCS System302™ by SMAR Ltd., was deployed in field
It is difficult to measure the heat release of combustion directly in level to implement field-based, distributed measurement and control.
practice. Therefore, an alternative measurement methodology is used FF authenticated intelligent fieldbus instruments and peripheral were
here. The measurement procedure is stated as follows: included in System302. FI302, IF302, LD302 and TT302 are the intel-
ligent transmitters and actuators used in the system. Control strategy
1. Measuring drum steam pressure and evaluating the changing rate configuration software Syscon 5.0 is also included in System302™.
of it, which is then multiplied by the capacity coefficient Cb;
2. Measuring main stream flow D; 4. Design and Implementation in Configuration
3. Heat release can be represented by the sum of D and the differential of
steam pressure in drum Pb with time. Configuration of control schemes was based on FF User Layer, which
is directly related to the process automation tasks themselves, and it
3. Control Implementation with Fieldbus System is based on distributed control or monitoring strategies of Function
Blocks. Function Blocks (FBs) are User Layer elements that encapsu-
The design objectives are to implement field-based and completely late basic automation functions and consequently make the configura-
distributed automation for the boiler combustion system with tion of a distributed industrial application modular and simplified.
fieldbus. It is well known that the fieldbus has many advantages Distributed among the field instruments, the FBs have their inputs
over DCS, and one of them is field-based control supported by and outputs linked to other blocks in order to perform distributed
function blocks. Some algorithms, which have to be implemented in closed control loop schemes. When blocks from different instruments
Main Control Unit (MCU) in DCS before can be achieved in field are linked together a remote link is configured and mapped to a cyclic
instruments now, which greatly improves the decentralization, safety message. The Foundation Fieldbus standardized a set of ten basic func-
and quality of control system [4]. tion blocks [5], a complementary set of eleven advanced control blocks,
The whole system was formulated into two levels: 1) supervisory and a special flexible function block intended to be fully configurable
level, 2) field-based measurement and control level. 10M Ethernet (i.e., internal logic and parameter) by users. The standard and advanced
is used in the communication of supervisory level while Foundation block sets provide mathematical and engineering calculations neces-
Fieldbus (FF) H1 (31.25Kb) is used in field level. Four redundant sary to configure typical industrial control loop strategies.
fieldbuses and bridges are installed with each fieldbus controlled by Foundation Fieldbus (FF) can be classified as a LAN (Local Area
a different bridge. Thererfore, the running of the whole system is not Network) for instruments used in process and industrial automation,
affected in case of breakdown from any fieldbus or bridge. Advanced with the ability to distribute the control application through a
supervisory and control configuration software was adopted in the network. The FF Data Link Layer supports two transmission policies:
upper level. Data exchange with field level, such as field data supervi- 1) scheduled cyclic data, and 2) sporadic (unscheduled) background
sion and setting point control etc, were implemented by means of data. These two communication policies share the physical bus, but
OPC. FF intelligent devices, such as transducers and actuators, were they are respectively segmented in cyclic time slots or periods. In the
deployed in field level. The structure of the whole control system scheduled communication period, most process variables generated
based on FF is shown in Figure 6. by periodic processes are transmitted cyclically according to a static
Fieldbus network and global schedule table loaded on the LAS node. This cyclic transmis-
Ethernet network, gateway sion mode has higher priority over acyclic transmission modes [6].
equipment, power supply, The general control strategies stated in Section 2.1, with the
HMI equipments (operator conventional cascade PID control as its backbone, were implemented
station) were all deployed in in field level by means of intelligent devices. The optimisation
redundancies. Operation of algorithm for the air-to-fuel ratio stated in Section 2.3 was deployed
the whole system will not be and executed in supervisory computer to make the evaluation
affected in case of fault in any easier. Necessary I/O data for optimisation computation, such as
single component. heat-release, airflow were acquired from the field devices and the
UPS was used in power optimised ratio sent back to them.
supply service. Enough power If no specific design is made to configuration of control strategies,
supply was ensured when data from field representing air and fuel, will be transmitted as
Figure 6: Structure and Framework of main power supply was in unscheduled background data when the computers require the data
the Control System Based on Foundation a fault state to secure the from the field instruments. This mechanism cannot meet the real
Fieldbus safe shutdown of the whole time requirement of control and data updating since time-critical

114 • Measurement + Control Vol 43/4 May 2010 www.instmc.org.uk

 Feature: A Boiler Combustion Control System with Combustion Ratio Optimisation and Soft Measurements

data exchange, which are also periodic as usual, cannot be guaran-

teed with unscheduled (aperiodic) communication.
To ensure a better control effect of combustion efficiency, the opti-
misation of air-to-fuel ratio need to be evaluated and transmitted to the
field instrument in real time.Therefore, data from field representing air
and fuel, need to be transmitted real time to the supervisory computer
as scheduled cyclic data.That means that some special designs are
needed to achieve this aim. In the related configuration, some additional
function blocks were added into the control configuration diagrams for
this purpose.Three additional function blocks, DFI-CONST, DFI-LLAG1
and DFI-LLAG2 were used in gateway DFI302 to exchange real time
data with field instruments by cyclic (scheduled) communication.
They are of constant and lead-lag type, respectively. DFI-LLAG1 and 1.Water feeding 2.Steam flow
DFI-LLAG2 were used to acquire data dealing with heat release and 3.Water level in drum 4.Furnace negative pressure
airflow by cyclic communication. DFI-CONST was used to the output
Figure 8: Plots of Controlled Parameters with the Designed System
air-to-fuel ratio optimized by supervisory computer to function block
PT-1805-RATIO in field instrument by cyclic communication. Output • Electricity consumption reduced by 63%
of function block PT-1802-ARTH-SUM, which represents heat release, • Profit increased by 100 million USD.
was connected to the input of DFI-LLAG1. Output of function block At the same time, advantages of using fieldbus, such as self-govern-
PT-1805-ARTH, which represents airflow, was connected to the input ing, danger-distribution and high reliability, were fully demonstrated
of DFI-LLAG2. Output of function block DFI-CONST, which represents by the system. Reliability, safety and economy during production in
the optimised air-to-fuel ratio, was connected to the input of PT-1805- the client enterprises were greatly improved.
RATIO.The complete configuration diagram of the control strategy
with gradient searching optimisation involved was shown in Figure 7, in 5. Conclusions
which 3 function blocks DFI-CONST, DFI-LLAG1 and DFI-LLAG2 with
specific purposes were highlighted in blue colour. Some methodologies and designs that were proved theoretically and
practically feasible and effective in measurement and control with fieldbus
technology for combustion control in boilers can be obtained as follows:
1. The general characteristic of differential pressure-air flow for an
air pre-heater is similar to that of the genuine orifice and bear the
similar nature as the orifice flow meter. Differential air pressure
between the inlet and outlet of air pre-heater can be selected as
air flow measurement by means of differential pressure detecting
instruments, which helps save cost and avoid throttling loss.
2. Changes in fuel can lead to change in heat release of boiler eventu-
ally. Heat release is an ultimate and reliable signal representing fuel.
Therefore, heat release of boiler can be taken as a measurement
to represent fuel actually.
Figure 7: Configuration Diagram of the Control Strategy with Gradient 3. Real time performance dealing with control and data exchange
Searching Optimisation over fieldbus network should be taken into full consideration.
Some special designs and techniques in systems configuration are
Optimisation of combustion, control of main steam pressure and important and requisite in improving the real time performance
furnace negative pressure were designed and implemented with the and control effects of the system. Simple function blocks can be
system using fieldbus. Automatic control was switched in when load chosen and added in configuration to help achieve data exchange
was above 50%. Following performance index were achieved when and control in real time over fieldbus.
load was less than 20%.
• Regulating time for combustion control system≤10min References
• Deviation in drum water level ≤±5mm 1. Caneiro A, Bonnat M, Fouletier J, Measurement and regulation of oxygen-content in
• Deviation main steam pressure ≤±0.05MPa selected gases using solid electrolyte cells, Journal of Applied Eletrochemistry, 1981.1, p83-90.
• Deviation in furnace negative pressure ≤±20Pa 2. Lu Yong, Xu Xiangdong, The Study of a New Method Incorporating the Soft Sensing of
Plots of several important parameters controlled by the designed Oxygen-content in Flue Gases, Journal of Engineering for Thermal Energy and Power, 2002.6,
system is shown in Figure 8. p614-617.
The system has run stably, with optimal combustion rate most 3. Jamais, S., Creaven, M., Calculated and measured oxygen contents in the gas flow
of the time, for two years in a large heat-supplying factory. It saved of a shrouded thermal spray process, Proceedings of the International Thermal Spray
energy consumption and reduced pollution greatly. According to Conference, 2001, p 903-910.
the economic benefit report by the client enterprise, the achieved 4. Chen J, Wang Z, Sun Y. How to improve control system performance using FF function
economic benefits were: blocks, IEEE international conference on control application, 2002, p 1022-1026.
• Number of dashboards reduced by 100% 5. Fieldbus Foundation, FF-890-1.3: Foundation specification function block application
• Number of control cabinets reduced by 50% process, part 1. Austin, USA. 1999.
• Coal consumption reduced by 20kg/t 6. Eduardo AndrÈ Mossin, Rodrigo Palucci Pantoni, A fieldbus simulator for training purposes,
• Cables and Wires reduced by 70% ISA Transactions, 2009.48, 132-141.

www.instmc.org.uk Measurement + Control Vol 43/4 May 2010 • 115