CN101859098A

CN101859098A - Boiler combustion rate self-optimizing method based on PID control policy

Info

Publication number: CN101859098A
Application number: CN 201010208604
Authority: CN
Inventors: 梁庚; 李文
Original assignee: North China Electric Power University
Current assignee: North China Electric Power University
Priority date: 2010-06-23
Filing date: 2010-06-23
Publication date: 2010-10-13
Anticipated expiration: 2030-06-23
Also published as: CN101859098B

Abstract

The invention uses the universal PID (Proportion Integration Differentiation) control policy, can realize the self-optimization of a boiler combustion 'fuel-air' rate value by selecting and designing PID controller input signals, has the characteristics of complete self adaptation, quick searching speed, good astringency and simple algorithm, solves the outstanding and typical problems in the traditional constant value rate method in boiler combustion control that the combustion rate can not be automatically regulated along with a working condition to keep optimal, and the like and also overcomes the defects in aspects of optimizing astringency and algorithm complexity in a combustion rate self-optimizing algorithm based on fixed step sizes and variable step sizes and realizability and easy possibility in a DCS (Distributed Control System) or FCS (Fly Control System) control system. The boiler combustion rate self-optimizing method has simple algorithm, is easy to realize, is very easy to realize on the traditional DCS/FCS platform for adopting the common PID control module and has better transportability.

Description

Boiler combustion rate self-optimizing method based on PID control strategy

Technical Field

The invention belongs to the technical field of automatic control, and relates to a boiler combustion rate self-optimizing control method based on a PID algorithm.

Background

The boiler plays an extremely important role in the energy industry, is widely applied to the industries of electric power, machinery, metallurgy, chemical industry, textile, paper making, food and the like, and plays an important role in industrial production and civil life. The combustion rate of the boiler is mainly reflected in the ratio of the fuel and the air quantity participating in combustion, the numerical value directly determines whether the heat energy generated in the combustion process can be fully utilized, and the method has extremely important significance for the economic operation, the energy conservation and the emission reduction of the boiler and the pollution reduction.

Typical existing boiler combustion control systems set the fuel-air ratio to a constant, even though the boiler combustion process remains operating at a fixed ratio of "fuel-air". However, many practical practices have shown that since boiler combustion is affected by various disturbance factors, especially for coal-fired boilers, the combustion efficiency is affected by various factors such as the type of coal and the quality of the coal, it is often difficult or even impossible to achieve an optimal combustion rate with a fixed "fuel-air" ratio, and even impossible to automatically maintain an optimal combustion rate with the variation of disturbance factors during control.

In some improved control schemes, the fuel quantity is represented by a furnace heat signal, which improves the problem brought by the fixed fuel-air ratio control scheme to a certain extent, but because the heat signal is acted by two main factors, namely fuel and air quantity, and is a multivariate function, the pure adoption of the heat signal as the fuel quantity has larger limitation and deficiency in accurately and objectively representing the actual fuel quantity, and the problems of automatic optimization of the boiler combustion rate and maintenance of the optimal combustion rate cannot be well solved.

In further research and application, scholars and engineering technicians design and realize a combustion rate self-optimization algorithm based on fixed step length and variable step length, and the basic idea is to judge which side of the relationship curve the current operating point is located on based on the relationship curve of air volume and combustion heat productivity by a trial and error method, and then adopt an algorithm of step-by-step approximation of fixed step length or variable step length to find the combustion rate optimal point located at the top point of the relationship curve. This self-optimizing algorithm overcomes the disadvantages of using a fixed "fuel-air" ratio control scheme, but its main problems are:

1. the determination of the optimizing step length mostly depends on experience;

2. the optimization time is long, and the optimization process is easy to oscillate near an optimal point;

3. the algorithm is complex, the calculation amount is large, and the realization of real-time control is not facilitated; meanwhile, because the mathematical operation and the logic judgment exist in the algorithm at the same time, the algorithm is difficult to realize by using an analog quantity process control basic module provided by the existing Distributed Control System (DCS) or field bus control system (FCS), and a user needs to adopt various programming languages related to the control system and the platform to realize the programming independently, so that the realizability, the usability and the portability of the algorithm based on the DCS/FCS control platform are greatly influenced.

Disclosure of Invention

Aiming at the problems of the combustion rate self-optimizing algorithm based on fixed step length and variable step length, the invention provides a combustion rate self-optimizing method based on a PID control strategy for combustion control of a boiler, which is characterized in that the method is realized by a universal PID controller module;

according to the operation logic of the PID control module or the function block, the input quantity e of the control module or the function block is enabled to be G (k) through proper configuration or parameter setting, namely:

e = G (k) = \frac{Q (k) - Q (k - 1)}{F (k) - F (k - 1)},

wherein, Q (k) is a heat signal sampling value at the current moment; q (k-1) is a heat signal sampling value at the last moment; f (k) is the sampling value of the air volume signal at the current moment; f (k-1) is the sampling value of the air volume signal at the previous moment;

at the next moment, the currently obtained sampling values Q (k), F (k) become Q (k-1) and F (k-1), new sampling values Q (k), F (k) are generated, G (k) is obtained through recalculation and is used as a process quantity input signal of the PID controller module, and a search value of the fuel-air ratio at the next moment is obtained through PID operation, so that a new search period is started;

the output OUT of the PID controller module is a search value of the fuel-air ratio at the current moment; after entering a steady state, the output of the controller is the optimal ratio of fuel to air;

the algorithm adopts different rules to carry out operation and search output according to the following conditions:

case 1: when the air volume detected in the two times is the same and the heating value is the same, the output value of the PD controller module, namely the ratio of fuel to air is not changed, and the state is equivalent to a steady state;

case 2: when the air volume detected in the two times is the same and the heating values detected in the two times are different, the output value of the PID controller module is kept unchanged;

case 3: the PID controller module performs output control until the output value is stable, which is equivalent to the searching process.

In the beginning of the whole search process, G (0) can be set to 0 or a small positive integer ε.

The air quantity signal sampling value can be replaced by an air quantity signal sampling value.

In the algorithm, the air volume F (k) ═ F (k-1) is not mathematically equal in a strict sense in engineering practice, and as long as | F (k) | < δ, F (k) | F (k-1) is considered to be F (k) | < δ, and δ is a certain dead zone value defined by a user.

The invention has the following beneficial effects:

1. the optimizing process is completely self-adaptive, and the searching step length and the searching speed automatically change along with the current optimizing state;

2. the search time is short, the search convergence is good, and the optimization precision is high;

3. the method has the advantages that the method is simple and easy to implement, is easy to implement on the existing DCS/FCS platform due to the adoption of a common PID control module, has good portability, is particularly suitable for the implementation of a field bus control system which has the characteristic of completely distributed control and is relatively simple in the provided field level control module, and can implement the real-time control of boiler combustion optimization on the basis of fully utilizing the technical advantages of the field bus due to the simple algorithm, high local operation speed and very small or basically no network flow data.

Drawings

The invention is described in detail below with reference to the attached drawing figures:

FIG. 1 is a graph showing a relationship between a calorific value Q of combustion of a boiler and an air volume (air volume) F at a constant fuel amount;

FIG. 2 is a graph showing a relationship between a calorific value Q and an air volume (air volume) F of boiler combustion;

FIG. 3 is a typical boiler combustion control system;

FIG. 4 is a boiler combustion control system based on a PID control strategy;

FIG. 5 shows the "fuel-air" optimum ratio R_opImplementing schemes of search algorithms;

fig. 6 shows a method (1) for implementing a PID control module with SP and PV inputs, where e is g (k);

fig. 7 shows a method (2) for implementing a PID control module with SP and PV inputs, where e is g (k);

fig. 8 shows a method for implementing e-g (k) by a PID control module without SP input;

FIG. 9 is a schematic diagram of a possible previous state point on the F-Q curve to the left of the peak;

FIG. 10 is the optimal path for the last state point from the top left;

FIG. 11 is the optimal path for the previous state point from the top right;

FIG. 12 is the optimal path for the last state point from the bottom left;

FIG. 13 is the optimal path for the previous state point from the bottom right;

FIG. 14 is a diagram of possible previous state points on the F-Q curve to the right of the peak.

Detailed Description

In a boiler system, when a unit fuel amount and an air volume (air volume) in which the unit fuel amount is mixed and burned at a ratio R are set, if a heat generation amount is a maximum value, a combustion rate of the boiler at that time is an optimum value, and a ratio R of "fuel-air" at that time is an optimum value R_op. This relationship can be described by an F-Q curve (air volume-heat curve) expressing the relationship between the air volume and the heat generation amount when the fuel amount is kept constant, as shown in fig. 1.

When the fuel quantity changes, a family of curves can be constructed on the basis of fig. 1, as shown in fig. 2. Each curve in the figure can express the relationship between the air volume and the heat productivity when the fuel quantity is a certain value. In the family of curves, the farther to the right the curve indicates the greater the fuel quantity, i.e. fuel quantity of curve 1 < fuel quantity of curve 2 < fuel quantity of curve 3.

FIG. 3 illustrates a typical boiler combustion control system in which the fuel-to-air ratio R is set at a constant set point to maintain the boiler combustion process operating at a fixed ratio of "fuel-to-air".

The boiler combustion control system based on the PID control strategy proposed by the invention is shown in FIG. 4. Here, the fuel-air ratio is no longer a fixed constant, but rather is determined by the "fuel-air" optimum ratio R_opAnd obtaining a search algorithm.

The search algorithm is further described below with reference to fig. 5:

1. a universal PID controller module is adopted;

2. according to the operation logic of the PID control module (or function block), the input quantity e of the module (or function block) is made to be g (k) by proper configuration or parameter setting, that is:

e = G (k) = \frac{Q (k) - Q (k - 1)}{F (k) - F (k - 1)}

wherein,

q (k): sampling value of heat signal at present moment;

q (k-1): sampling value of heat signal at last moment;

f (k): sampling value of air volume (air volume) signal at current moment;

f (k-1): sampling value of air quantity (air quantity) signal at last moment;

for example, in a control strategy configuration of a Distributed Control System (DCS) or a field bus control system (FCS), the implementation method of e ═ g (k) may be:

(1) for a PID control block (or functional block) having SP and PV inputs, if the arithmetic logic of the block is defined as e-SP-PV, then g (k) is connected to the SP terminal and the input terminal PV is 0; otherwise, if the arithmetic logic is defined as e-PV-SP, g (k) is connected to the PV end, and the input SP is 0.

As shown in fig. 6 and 7.

(2) For a PID module that does not have an SP input, i.e., the set point value is set by the user by means of an internal parameter of the function block (e.g., in some FCS configuration software), the operating logic of the PID module is set to e-PV-SP, g (k) is then connected to the PV end, and the internal parameter SP is set to 0. As shown in fig. 8.

The output OUT of the PID controller module is the search value R (k) of the "fuel-air" ratio R at the current moment (current step); after entering the steady state, the output of the controller is the optimal ratio R of' fuel-air_op；

4. At the next moment, the currently obtained sampling values Q (k), F (k) become Q (k-1) and F (k-1), new sampling values Q (k), F (k) are generated, G (k) is obtained through recalculation and is used as a process quantity input signal of a PID controller module, and a search value R (k) of the fuel-air ratio R at the next moment is obtained through PID operation, so that a new search period is started;

5. g (0) may be set to 0 or a small positive integer ε at the beginning of the entire search process;

6. the algorithm adopts different rules to carry out operation and search output according to the following conditions:

case 1: when the air volume detected in the two previous times and the air volume detected in the two previous times are the same, namely F (k) ═ F (k-1), and the calorific value is the same, namely Q (k) ═ Q (k-1), the PID controller output value, namely the ratio R of 'fuel-air' is not changed, and the state is equivalent to a steady state;

and 2, when the air volume detected in the last two times is the same, namely F (k) ≠ F (k-1), and the heat productivity detected in the last two times is different, namely Q (k) ≠ Q (k-1), R output by the PID controller is also kept unchanged. Because:

the factors causing the variation of the calorific value are: 1) change in fuel quantity 2) other disturbance factors than the fuel quantity (i.e., when the fuel quantity is constant, and some other minor disturbance factors are changed, the heating value is also changed). If the amount of heat generation is changed due to a change in the amount of fuel, as is clear from the control strategy shown in fig. 3, the air volume is changed later, and the "same air volume" detected at this time, that is, F (k) ═ F (k-1), is only temporary. Due to the change of the fuel quantity, the change of the air quantity is detected quickly, namely F (k) is not equal to F (k-1), and processing and searching output are performed according to the rule corresponding to the 'case 3'. If the heat value is changed due to other disturbance factors except the fuel quantity, the design of the controller is designed according to an F-Q curve (a relation curve of air volume and heat value when the fuel quantity is constant) as a basic basis, so that in the case, the controller adopts a conservative strategy, namely the controller keeps the output R unchanged, because the influence rule of the air volume change on the heat value when other disturbance factors act is not known definitely.

And 3, when the air volume detected at the last two times is different and the calorific value is different, namely F (k) ≠ F (k-1) and Q (k) ≠ Q (k-1), the output of the PID controller, namely the R value of the ratio of fuel to air, is subjected to output control according to the strategy as described in the figure 5, which is equivalent to the process of searching.

7. Further, it should be noted that F (k) ═ F (k-1) in the algorithm is not mathematically equivalent in a strict sense in engineering practice, and is considered as F (k) ═ F (k-1) | < δ (δ is a certain dead zone value defined by a user) as long as | F (k) ═ F (k-1) is satisfied.

The search process for case 3 above can be further explained as:

as shown in fig. 9, the points { Q (k) }, F (k) } on the F-Q curve represent the current state, which can only be changed from the previous state { Q (k-1), F (k-1) }, characterized by four directions (i.e., up-left, down-left, up-right, down-right, all possible previous states always coming from one of the four directions. The last state may come from the same F-Q curve or from other F-Q curves on the family of curves (corresponding to different fuel quantities).

The relationship between the previous state { Q (k-1), F (k-1) } in these four directions and the current state { Q (k), F (k) }, will be discussed one by one, and the case where the current point is on the left side of the curve peak will be detailed here, and the same analysis will be applied to the case where the current point is on the right side of the curve peak.

1. The previous state comes from direction (left-up direction): in this direction, the relationship between the previous state and the current state is: q (k) -Q (k-1) < 0, F (k) -F (k-1) > 0, namely the air volume is increased, but the unit heating value is reduced. Therefore, based on the input from the optimizing PID controller:

\frac{Q (K) - Q (K - 1)}{F (K) - F (K - 1)} < 0

therefore, the output from the optimizing PID controller will decrease, i.e., the ratio R decreases, and thus the air volume will decrease, and as the air volume decreases, the heating value will decrease, i.e., the current point will move down the curve to the point { Q (k +1), F (k +1) }, as shown in fig. 10, assuming that the fuel amount is constant at this time, as per the above combustion curve. After moving to point { Q (k +1), F (k +1) }, compared to the previous state point { Q (k), F (k) }: q (k +1) -Q (k) < 0, F (k +1) -F (k) < 0, where the inputs from the optimizing PID controller are:

\frac{Q (K + 1) - Q (K)}{F (K + 1) - F (K)} > 0

the output from the optimizing PID controller will increase, i.e., the ratio R increases, and the air volume will increase, i.e., begin to go up the F-Q curve, moving toward the next state point { Q (k +2), F (k +2) }. Therefore, after the optimization direction is adjusted by at most one step, the correct optimization direction is entered again.

In summary, the optimal route is shown in fig. 10, wherein 1, 2, and 3 indicate the optimal route.

2. The previous state comes from direction (upper right direction): the self-optimizing principle is the same as the analysis process, and the optimizing route is shown in fig. 11.

3. The previous state comes from direction (left-down direction): the optimization route is shown in fig. 12.

4. The previous state comes from direction (lower right direction): the optimization route is shown in fig. 13.

For the case where the current point is to the right of the peak of the F-Q curve, as shown in FIG. 14, the analysis process and operating principles are the same as before.

It can be seen that by proper selection and design of the signal form of the controller input, self-optimization of the "fuel-air" ratio can be achieved using conventional PID control strategies.

The invention has been tried in the combustion control based on DCS and FCS of a plurality of domestic thermal power plants and heating boilers, the system has been stably operated for more than 1 year since the field implementation and the operation, the fuel consumption is saved to a great extent, the emission is reduced, the coal consumption of each ton of steam is reduced by 20Kg according to the statistical data provided by a certain heating enterprise, and compared with the combustion control system before the patent is implemented, the coal consumption is reduced by 30%. Meanwhile, due to strong usability and realizability, the invention can realize a fully-distributed measurement and control system based on an advanced field bus technology, greatly improve the comprehensive automation level of a plant level, greatly reduce the cost of manpower and system maintenance, and increase the direct and indirect economic benefits of an enterprise by 500 ten thousand yuan per year.

Claims

1. A boiler combustion rate self-optimizing method based on PID control strategy is characterized in that the method is realized by a universal PID controller module;

e = G (k) = \frac{Q (k) - Q (k - 1)}{F (k) - F (k - 1)},

case 1: when the air volume detected in the two times is the same and the heating value is the same, the output value of the PID controller module, namely the ratio of fuel to air is not changed, and the stable state is equivalent to the stable state;

2. The method of claim 1, wherein G (0) is set to 0 or a small positive integer e at the beginning of the entire search process.

3. The method of claim 1, wherein the air volume signal samples are replaced with air volume signal samples.

4. The method according to claim 1, wherein the air volume F (k) ═ F (k-1) in the algorithm is not mathematically equal in a strict sense in engineering practice, and as long as | F (k) | < δ is satisfied, F (k) | < δ is considered as F (k) ═ F (k-1), and δ is a certain dead zone value defined by a user.