CN110618706A - Multistage intelligent denitration online optimization control system based on data driving - Google Patents
Multistage intelligent denitration online optimization control system based on data driving Download PDFInfo
- Publication number
- CN110618706A CN110618706A CN201910921188.6A CN201910921188A CN110618706A CN 110618706 A CN110618706 A CN 110618706A CN 201910921188 A CN201910921188 A CN 201910921188A CN 110618706 A CN110618706 A CN 110618706A
- Authority
- CN
- China
- Prior art keywords
- value
- ammonia
- nox
- concentration
- boiler
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B01—PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
- B01D—SEPARATION
- B01D53/00—Separation of gases or vapours; Recovering vapours of volatile solvents from gases; Chemical or biological purification of waste gases, e.g. engine exhaust gases, smoke, fumes, flue gases, aerosols
- B01D53/34—Chemical or biological purification of waste gases
- B01D53/74—General processes for purification of waste gases; Apparatus or devices specially adapted therefor
- B01D53/86—Catalytic processes
- B01D53/8621—Removing nitrogen compounds
- B01D53/8625—Nitrogen oxides
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B01—PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
- B01D—SEPARATION
- B01D53/00—Separation of gases or vapours; Recovering vapours of volatile solvents from gases; Chemical or biological purification of waste gases, e.g. engine exhaust gases, smoke, fumes, flue gases, aerosols
- B01D53/34—Chemical or biological purification of waste gases
- B01D53/74—General processes for purification of waste gases; Apparatus or devices specially adapted therefor
- B01D53/86—Catalytic processes
- B01D53/90—Injecting reactants
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D7/00—Control of flow
- G05D7/06—Control of flow characterised by the use of electric means
- G05D7/0617—Control of flow characterised by the use of electric means specially adapted for fluid materials
- G05D7/0629—Control of flow characterised by the use of electric means specially adapted for fluid materials characterised by the type of regulator means
- G05D7/0635—Control of flow characterised by the use of electric means specially adapted for fluid materials characterised by the type of regulator means by action on throttling means
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Abstract
The invention relates to a data-driven multi-stage intelligent denitration online optimization control system, which is used for controlling ammonia escape to be a minimum value on the premise of ensuring that NOx emission does not exceed the standard.
Description
Technical Field
The invention relates to the technical field of automatic control of thermal power plants, in particular to a multistage intelligent denitration online optimization control system based on data driving.
Background
At present, most power plants complete the transformation work of ultra-low emission, and the NOx emission concentration must be controlled to be 50mg/Nm according to relevant environmental protection regulations3Below the range, the NOx emission index will be controlled to 35mg/Nm according to the trend of development3Within the range.
Due to various factors such as large delay and large inertia of objects in the denitration process, low representativeness of measuring points in the operation process, small mediation allowance and the like, the denitration automatic adjustment quality of the coal-fired power plant is poor.
In the denitration control, the ammonia injection amount is insufficient, the NOx exceeds the standard, and the environmental protection assessment value is unqualified; if the ammonia injection is excessive, the ammonia slip is large, NH3Attached to the surface of the catalyst, blocks catalyst pore channels, causes catalyst inactivation and blockage of an air preheater, is dispersed in flue gas to corrode pipelines and downstream equipment, and harms the operation safety of a unit.
At present, after denitration ammonia injection of a unit is automatically put into operation, NOx fluctuation is large, oscillation is not easy to be stable, and especially when a load is changed, a coal mill is started and stopped, and coal blending combustion is carried out, NOx dynamic deviation is large, so that the development requirements of intellectualization, informatization and digitization of a power plant cannot be met.
Disclosure of Invention
In view of the above situation, in order to overcome the defects of the prior art, the present invention aims to provide a data-driven online optimization control system for multi-stage intelligent denitration, so as to control the ammonia escape to a minimum value on the premise of no exceeding of NOx emission.
The technical scheme of the invention is as follows:
a data-driven multi-stage intelligent denitration online optimization control system comprises five control stages, namely a rapid protection stage, an intelligent feedforward stage, an accurate control stage, an ammonia escape control stage and an online optimization stage, wherein an ammonia flow deviation value of the total ammonia demand and the actual ammonia flow is obtained through the five control stages, and the opening degree of an ammonia flow regulating valve is correspondingly regulated according to the ammonia flow deviation value, so that the ammonia escape is controlled at the lowest value on the premise that NOx emission does not exceed the standard;
wherein:
fast protection level: calculating ammonia demand through the concentration value and the change rate of the clean flue gas NOx to obtain primary ammonia demand, and rapidly increasing ammonia to prevent the over-standard phenomenon caused by sudden rising of the concentration of the NOx so as to control the NOx emission not to exceed the standard;
intelligent feed-forward stage: the method comprises the steps of outputting secondary ammonia demand through a feedforward control module formed by a linear variable parameter model structure by taking NOx concentration at an SCR inlet of a boiler A/B side, unit load, coal quantity, air quantity, measured value of SCR outlet concentration of the boiler A/B side and NOx concentration setting guidance value of the SCR outlet of the boiler A/B side as input variables, and directly changing ammonia flow when any variable changes so as to realize the rapidity of NOx emission control;
accurate control level: a set value prediction control loop and a cascade negative feedback prediction control loop;
the set value prediction control loop comprises a NOx concentration calculation module formed by a linear variable parameter model structure, a net flue gas NOx concentration set value is used as an input variable, the output is a boiler A/B side SCR outlet NOx concentration setting guide value, the boiler A/B side SCR outlet NOx concentration setting guide value is used as one variable of a feedforward control module in an intelligent feedforward level, the concentration setting guide value and a boiler A/B side SCR outlet NOx concentration correction value output by an ammonia escape control level are combined to obtain a boiler A/B side SCR outlet NOx concentration set value, the concentration set value is compared with a boiler A/B side SCR outlet NOx concentration measurement value to obtain a deviation value of the boiler A/B side SCR outlet NOx concentration set value and the measurement value, and the deviation value is used as an input quantity of a cascade negative feedback prediction control loop;
the cascade negative feedback prediction control loop comprises a prediction controller based on a compact generalized prediction model, the input of the prediction controller is the deviation value of a NOx concentration set value and a measured value at the outlet of an SCR (selective catalytic reduction) at the A/B side of a boiler, the output of the prediction controller is ammonia gas three-level demand, the nitrogen oxide value at the outlet of a denitration reactor under the current ammonia injection amount can be obtained by calculating in advance through the prediction model, whether the ammonia injection amount is proper or not is judged in advance, and then an adjustment suggested value of the ammonia injection amount is given;
summing the primary ammonia demand, the secondary ammonia demand and the tertiary ammonia demand to obtain the total ammonia demand;
transmitting the ammonia flow deviation value of the total ammonia demand and the actual ammonia flow to a PID (proportion integration differentiation) controller of a denitration system of the power plant for opening degree regulation control of an ammonia flow regulating valve;
ammonia slip control stage: correcting the NOx concentration setting guide value at the SCR outlet of the A/B side of the boiler according to the ammonia escape measured value to obtain the corrected value of the NOx concentration at the SCR outlet of the A/B side of the boiler, and improving the NOx emission concentration value at the SCR outlet to achieve the purpose of reducing ammonia escape;
and (3) online optimization stage: and updating the parameters of the rest four control levels on line through a particle swarm optimization algorithm according to historical operating data.
Preferably, the starting conditions of the fast protection stage are as follows:
a. net flue gas NOx concentration value>47mg/Nm3(ii) a b. Rate of change of net flue gas NOx concentration value>5mg/Nm3/min；
When the two conditions of a and b are met simultaneously, the first-level demand of ammonia gas is output, and the output value is as follows:
ammonia gas current flow value x (net flue gas NOx current concentration value-47) x 5% x protection action amplitude coefficient
The over-standard phenomenon caused by the sudden rise of the concentration of the NOx can be prevented, and the NOx emission is controlled not to be over-standard.
Preferably, the fast protection level has the highest priority.
Preferably, the feedforward control module of the intelligent feedforward stage has a specific structure that:
assuming that the input of the system at time t is u (t), the output is y (t), I and J are the orders of the model autoregressive part and the moving average part respectively, and ξ (t) is white noise, the linear parameter change model of the system can be expressed as:
wherein p, q is 1,2, …, n; 1,2, … I1；r＝1,2,…,m；j＝1,2,…,Jr(ii) a S is 1,2, …, S; m is the number of input variables, n is the number of output variables, and S is the number of selected reference working conditions;andin order to input the weight coefficient,is an exponential factor polynomial and is used as a reference,andin order to weight the offset, the weight is,is a function ofThe central point of (2), namely the selected reference working condition point; delta is a scaling factor, delta > 0.
In the control system, the number m of input variables is 6, the input variables are respectively the concentration u (1) of NOx at the SCR inlet of the A/B side of a boiler, the unit load u (2), the coal quantity u (3), the air quantity u (4), the measured value u (5) of the concentration of the SCR outlet of the A/B side of the boiler and the set guidance value u (6) of the concentration of the NOx at the SCR outlet of the A/B side, the number n of output variables is 1, the output y (1) is the secondary demand of ammonia gas, the reference working condition number S is 7, x is the time-varying parameter of the system, and the unit load is selected as the time-varying parameter;
preferably, in the cascade negative feedback prediction control loop, the prediction controller based on the compact generalized prediction model has a structure that:
y(k+j)＝Lj(q-1)Δu(k-1)+y0(k+j)
wherein: j ═ 1, Λ, P; p is the predicted step number; p is the predicted step number; k is the current time; l is the number of items in the control domain; u is the system input; y0 is the contribution of past inputs and outputs to future time output
The nitrogen oxide value at the outlet of the denitration reactor under the current ammonia injection amount can be calculated in advance through a prediction model;
preferably, the starting conditions of the ammonia slip control stage are:
a. net flue gas NOx concentration<40mg/Nm3；
b. The set value of the ammonia injection valve is not changed within 10 min;
c. the average value of ammonia escape is more than 10 ppm;
when the three conditions a, B and c are met simultaneously, overlapping the corrected value of the concentration of the NOx at the SCR outlet of the A/B side of the boiler into the set guidance value of the concentration of the NOx at the SCR outlet of the A/B side of the boiler every 15min to achieve the purpose of improving the concentration value of the NOx, thereby reducing the ammonia escape;
after the ammonia escape control level intervenes in automatic control, if the net flue gas NOx concentration value is more than 45mg/Nm in 10min and the average value is more than 45mg/Nm3The correction values are not superimposed.
In the five control levels, the rapid protection level calculates the ammonia demand according to the concentration value and the change rate of the NOx in the clean flue gas to obtain the primary demand of the ammonia gas, so that the ammonia gas can be rapidly increased, and the overproof phenomenon caused by the sudden rising of the concentration of the NOx is prevented, thereby controlling the NOx emission not to be overproof; the intelligent feedforward stage takes the concentration of NOx at an inlet of an SCR at the A/B side of a boiler, the load of a unit, the coal quantity, the air quantity, the measured value of the concentration of an SCR outlet at the A/B side of the boiler and the set guidance value of the concentration of NOx at the outlet of the SCR at the A/B side of the boiler as input variables to output the secondary demand quantity of ammonia gas, and when any variable is changed, the flow of ammonia gas is directly changed so as to realize the rapidity of NOx emission control; the method comprises the steps that three levels of ammonia demand are obtained through a precise control level, the total ammonia demand is finally obtained, and the ammonia flow deviation value of the total ammonia demand and the actual ammonia flow is transmitted to a PID (proportion integration differentiation) controller of a power plant denitration system to carry out opening degree regulation control on an ammonia flow regulating valve; on the premise of not exceeding the standard of NOx emission, ammonia escape is controlled to be the lowest value.
Drawings
FIG. 1 is a schematic diagram of the multi-stage intelligent denitration online optimization control system, wherein the labeled meanings in the diagram are shown in the following table:
FIG. 2 is a flow chart of the particle swarm optimization algorithm (PSO) in the online optimization stage of the present invention.
Detailed Description
The following examples further illustrate the embodiments of the present invention in detail.
The method comprises five control levels, namely a rapid protection level, an intelligent feedforward level, an accurate control level, an ammonia escape control level and an online optimization level, wherein an ammonia flow deviation value of the total ammonia demand and the actual ammonia flow is obtained through the five control levels, and the opening degree of an ammonia flow regulating valve is correspondingly regulated according to the ammonia flow deviation value, so that the ammonia escape is controlled to be the lowest value on the premise that the NOx emission does not exceed the standard;
wherein:
fast protection level: calculating ammonia demand through the concentration value and the change rate of the clean flue gas NOx to obtain primary ammonia demand, and rapidly increasing ammonia to prevent the over-standard phenomenon caused by sudden rising of the concentration of the NOx so as to control the NOx emission not to exceed the standard;
intelligent feed-forward stage: the method comprises the steps that a NOx concentration at an inlet of an SCR at a boiler A/B side, unit load, coal quantity, air quantity, a measured value of the SCR outlet concentration at the boiler A/B side and a set guidance value of the NOx outlet concentration at the boiler A/B side are used as input variables, a feedforward control module formed by a linear parameter modeling (LPV) is used for outputting secondary demand of ammonia gas, and when any variable changes, the flow of the ammonia gas is directly changed to realize the rapidity of NOx emission control;
accurate control level: a set value prediction control loop and a cascade negative feedback prediction control loop;
wherein the set value prediction control loop comprises a linear parameter varying model structure (LPV, namely a network model 2 in a schematic diagram) to form a NOx concentration calculation module, the set value of the concentration of the NOx in the clean flue gas is taken as an input variable, the output is a set guide value of the concentration of the NOx at the outlet of the SCR at the A/B side of the boiler, the boiler A/B side SCR outlet NOx concentration setting guide value is used as one variable of a feedforward control module in an intelligent feedforward stage, simultaneously, the concentration setting guide value and a corrected value of the NOx concentration at the outlet of the SCR at the A/B side of the boiler output by the ammonia escape control stage are summed to obtain a set value of the NOx concentration at the outlet of the SCR at the A/B side of the boiler, the set value of the concentration is compared with a measured value of the NOx concentration at the outlet of the SCR at the A/B side of the boiler to obtain a deviation value of the set value and the measured value of the NOx concentration at the outlet of the SCR at the A/B side of the boiler, and the deviation value is used as;
the cascade negative feedback prediction control loop comprises a prediction controller based on a compact generalized prediction model, the input of the prediction controller is the deviation value of a NOx concentration set value and a measured value at the outlet of an SCR (selective catalytic reduction) at the A/B side of a boiler, the output of the prediction controller is ammonia gas three-level demand, the nitrogen oxide value at the outlet of a denitration reactor under the current ammonia injection amount can be obtained by calculating in advance through the prediction model, whether the ammonia injection amount is proper or not is judged in advance, and then an adjustment suggested value of the ammonia injection amount is given;
summing the primary ammonia demand, the secondary ammonia demand and the tertiary ammonia demand to obtain the total ammonia demand;
transmitting the ammonia flow deviation value of the total ammonia demand and the actual ammonia flow to a PID (proportion integration differentiation) controller (an inner loop controller) of a denitration system of the power plant for opening degree regulation control of an ammonia flow regulating valve;
ammonia slip control stage: correcting the NOx concentration setting guide value at the SCR outlet of the A/B side of the boiler according to the ammonia escape measured value to obtain the corrected value of the NOx concentration at the SCR outlet of the A/B side of the boiler, and improving the NOx emission concentration value at the SCR outlet to achieve the purpose of reducing ammonia escape;
and (3) online optimization stage: and updating the parameters of the rest four control levels on line through a particle swarm optimization algorithm according to historical operating data.
The starting conditions of the fast protection stage are as follows:
a. net flue gas NOx concentration value>47mg/Nm3(ii) a b. Rate of change of net flue gas NOx concentration value>5mg/Nm3/min；
When the two conditions of a and b are met simultaneously, the first-level demand of ammonia gas is output, and the output value is as follows:
ammonia gas current flow value x (net flue gas NOx current concentration value-47) x 5% x protection action amplitude coefficient
The over-standard phenomenon caused by the sudden rise of the concentration of the NOx can be prevented, and the NOx emission is controlled not to be over-standard.
The quick protection stage comprises a clean flue gas NOx concentration protection control module C (NOx) and is used for collecting a clean flue gas NOx concentration value and the change rate of the clean flue gas NOx concentration value, judging the starting condition of the quick protection stage, and sending the output primary demand of the ammonia gas to an addition block ADD to sum with the secondary demand of the ammonia gas.
The concentration of the clean flue gas NOx is the concentration of the NOx behind the desulfurizing tower and the concentration of the NOx at the inlet of the chimney, and the concentration value and the change rate of the concentration value can be directly collected by an original denitration system of a power plant.
The protection action amplitude coefficient is obtained through online optimization level calculation, and the value range is generally 0.6-1.5.
The fast protection level has the highest priority.
The intelligent feedforward level feedforward control module has the specific structure that:
assuming that the input of the system at time t is u (t), the output is y (t), I and J are the orders of the model autoregressive part and the moving average part respectively, and ξ (t) is white noise, the linear parameter change model of the system can be expressed as:
wherein p, q is 1,2, …, n; 1,2, … I1；r＝1,2,…,m；j＝1,2,…,Jr(ii) a S is 1,2, …, S; m is the number of input variables, n is the number of output variables, and S is the number of selected reference working conditions;andin order to input the weight coefficient,is an exponential factor polynomial and is used as a reference,andin order to weight the offset, the weight is,is a function ofThe central point of (2), namely the selected reference working condition point; delta is a scaling factor, delta > 0.
In the control system, the number m of input variables is 6, the input variables are respectively the concentration u (1) of NOx at the SCR inlet of the A/B side of a boiler, the unit load u (2), the coal quantity u (3), the air quantity u (4), the measured value u (5) of the concentration of the SCR outlet of the A/B side of the boiler and the set guidance value u (6) of the concentration of the NOx at the SCR outlet of the A/B side, the number n of output variables is 1, the output y (1) is the secondary demand of ammonia gas, the reference working condition number S is 7, x is the time-varying parameter of the system, and the unit load is selected as the time-varying parameter;
the time-varying parameters in the linear variable parameter model directly determine the time-varying characteristics of the parameters in the system equation, one system is selected to be measurable in the mechanism analysis process of a modeling object, the time variable closely related to the change of the model parameters is particularly important, whether the selection of the time-varying parameter x is proper or not is related to the rationality of the time-varying parameter model of the whole system, and for the denitration ammonia injection control system, because the unit load is corresponding to the fuel quantity and the air quantity, and the system variable is easily measured, the time-varying parameters are set as the unit load (the current load) in the feedforward control model.
The NOx concentration of an SCR inlet at the A/B side of the boiler, unit load, coal quantity, air quantity and the measured value of the SCR outlet concentration at the A/B side of the boiler are directly collected by an original denitration system of a power plant, and the NOx concentration setting guide value at the SCR outlet at the A/B side of the boiler is output by a net flue gas NOx set value prediction control loop in a precise control stage.
Weight coefficientAndweight biasAndthe scaling factor δ is calculated by an online optimization stage, δ typically floating around 0.75.
And after the feedforward control module outputs the secondary ammonia demand, the secondary ammonia demand is sent to an addition block ADD to be summed with the primary ammonia demand.
The method comprises the steps that a NOx concentration calculation module in a set value prediction control loop and a feedforward control module in an intelligent feedforward level both adopt linear variable parameter model structures, so that the model structures are the same, and the difference is that the number m of input variables of the NOx concentration calculation module is 1, the input variables are set values of the concentration of clean flue gas NOx, the number n of output variables is 1, the output variables are set guide values of the concentration of NOx at an SCR outlet at the A/B side of a boiler, the number S of reference working conditions is 7, and the set values of the concentration of the clean flue gas NOx are selected as time-varying parameters;
the concentration set value of the clean flue gas NOx is 25-45 generally, and the national environmental protection regulation 50 exceeds the standard.
Weight coefficientAndweight biasAndthe scaling coefficient delta is obtained through online optimization level calculation;
in the cascade negative feedback prediction control loop, the structure of a prediction controller based on a compact generalized prediction model is as follows:
y(k+j)＝Lj(q-1)Δu(k-1)+y0(k+j)
wherein: j ═ 1, Λ, P; p is the predicted step number; k is the current time; l is the number of items in the control domain; u is the system input; y is0Contributions of past inputs and outputs to outputs at a future time; the nitrogen oxide value at the outlet of the denitration reactor under the current ammonia injection amount can be calculated in advance through a prediction model.
The input of the system k at the moment is set as a deviation value u (t) of a NOx concentration set value and a measured value at the outlet of an SCR at the A/B side of the boiler, and the output is three-level ammonia demand y (t), wherein:
A(q-1)y(k)＝B(q-1)u(k-1) (10)
the pair of equations (10) is modified:
the factor can be divided by a long divisionDecomposing is carried out, if j items before decomposition are present
In the formula:in order to obtain the quotient polynomial,is a remainder polynomial, j is 1, Λ P.
Equation (12) is actually the Diophantine equation.
By substituting formula (12) for formula (11), and by using formula (10):
y(k)＝Ej(q-1)B(q-1)Δu(k-1)+q-jFjy(k) (13)
the predicted output at time j in the future may be expressed as:
y(k+j)＝qjEj(q-1)B(q-1)Δu(k-1)+Fjy(k) (14)
using "1" as a dividing polynomial to divide Ej(q-1)B(q-1) The decomposition by long division can obtain:
wherein:corresponding to present and future control input portions;
corresponding to the past control input section.
Then there are:
y(k+j)＝qjGj(q-1)Δu(k-1)+Hj(q-1)Δu(k-1)+Fjy(k) (16)
wherein: j ═ 1, Λ, P; p is the number of predicted steps.
As can be seen from equation (16): the key of the output of the system for predicting the future time is to determine an operator polynomial Gj(q-1)、Hj(q-1) And Fj(q-1) The method of (3).
For determining the coefficients G in the prediction relation (16)j(q-1)、Hj(q-1) And Fj(q-1) The following method can be adopted:
according to formula (12):
equation (17) may be changed to:
Fj(q-1)＝qj[Ej+1(q-1)-Ej(q-1)]A(q-1)Δ+q-1Fj+1(q-1) (18)
because Ej(q-1)、Ej+1(q-1) Respectively, the quotient polynomials of degree j and (j +1), so:
ej+1,j＝qj[Ej+1(q-1)-Ej(q-1)] (19)
and because of a0Q on both sides of formula (18) is compared as 10The secondary coefficient is obtained:
ej+1,j＝fj0 (20)
formula (20) is substituted for formula (19):
fj0＝qj[Ej+1(q-1)-Ej(q-1)] (21)
formula (21) is obtained by substituting formula (18):
Fj+1(q-1)＝q[Fj(q-1)-A(q-1)Δfj0]
or Fj+1(q-1)＝q[Fj(q-1)-fj0]-fj0{q[A(q-1)-1]-A(q-1)}
Substituting subscript j for subscript (j +1) in the above formula yields Fj(q-1) The recursion of (1) is:
Fj(q-1)＝q[Fj-1(q-1)-fj-1,0]-fj-1,0{q[A(q-1)-1]-A(q-1)} (22)
multiplying B (q) on both sides of equation (21)-1) And is obtained by using the formula (15):
fj0B(q-1)+Hj(q-1)＝qj[Gj+1(q-1)-Gj(q-1)]+q-1Hj+1(q-1) (23)
due to Gj(q-1)、Gj+1(q-1) Are respectively Ej(q-1)、B(q-1) Regarding the j-th and (j +1) -th order quotient polynomials of "1" polynomial, the first term on the right of the equation ((23) is constant, i.e.:
gj+1,j＝qj[Gj+1(q-1)-Gj(q-1)] (24)
let the left side of equation (23) equal toNamely, it is
Wj(q-1)＝fj0B(q-1)+Hj(q-1) (25)
Formula (24) and formula (25) are substituted for formula (23):
Wj(q-1)＝gj+1,j+q-1Hj+1(q-1) (26)
two sides q of comparative formula (26)0The coefficient is obtained:
gj+1,j＝wj0 (27)
formula (27) is arranged in place of formula (26):
Hj+1(q-1)＝q[Wj(q-1)-wj0] (28)
substitution of subscript j in formula (25) with subscript (j-1)Obtaining Wj-1(q-1) The recursion of (1) is:
Wj-1(q-1)＝fj-1,0B(q-1)+Hj-1(q-1) (29)
substitution of subscript j for subscript (j +1) in formula (28) with subscript j, provides Hj(q-1) The recursion of (1) is:
Hj(q-1)＝q[Wj-1(q-1)-wj-1,0] (30)
equation (24) can be written as:
Gj+1(q-1)＝Gj(q-1)+q-jwj,0 (31)
by substituting subscript j for subscript (j +1) in the above formula, G can be obtainedj(q-1) The recursion of (1) is:
Gj(q-1)＝Gj-1(q-1)+q-(j-1)wj-1,0 (32)
or
Comparative formulas (33) andtherefore, the following steps are carried out:
gji＝wi-1,0 (34)
equation (16) can be rewritten as:
y(k+j)＝[qjGj(q-1)]Δu(k-1)+y0(k+j) (35)
in the formula: y is0(k+j)＝Hj(q-1)Δu(k-1)+Fj(q-1) y (k) is the contribution of the past inputs and outputs to the output at the future time (k + j), j ═ 1, Λ P.
The starting conditions of the recurrence formula are: g0(q-1)＝0，H0(q-1)＝0，F0(q-1)＝1。
The first term coefficient polynomial to the right of the equal sign of equation (35) can be determined by the following method:
if the number of selected control field entries is M, equation (36) is rewritten as:
thus, equation (35) can be rewritten as:
y(k+j)＝Lj(q-1)Δu(k-1)+y0(k+j) (38)
wherein: j ═ 1, Λ P.
Equation (38) is the relationship used by the closed-loop prediction module to predict the system output at the future time.
The starting conditions of the ammonia escape control stage are as follows:
a. net flue gas NOx concentration<40mg/Nm3；
b. The set value of the ammonia injection valve is not changed within 10 min; the ammonia injection damper set point is a (required, desired) quantity given by the operating operator;
c. the average value of ammonia escape is more than 10 ppm;
when the three conditions a, B and c are met simultaneously, overlapping the corrected value of the concentration of the NOx at the SCR outlet of the A/B side of the boiler into the set guidance value of the concentration of the NOx at the SCR outlet of the A/B side of the boiler every 15min to achieve the purpose of improving the concentration value of the NOx, thereby reducing the ammonia escape;
after the ammonia escape control level intervenes in automatic control, if the net flue gas NOx concentration value is more than 45mg/Nm in 10min and the average value is more than 45mg/Nm3No correction value is superposed;
if the following occurs:
1) the operator can also manually cut off the automatic control;
2) the operator modifies the set value of the ammonia injection valve;
the boiler a/B side SCR outlet NOx concentration correction value is reset to zero.
The ammonia escape control stage comprises an ammonia escape controller C (NH3) which is used for collecting a clean flue gas NOx concentration value, an ammonia injection valve set value and an ammonia escape average value, judging the starting condition of the ammonia escape control stage, and sending an output corrected value to an addition block ADD to sum with a boiler A/B side SCR outlet NOx concentration setting guide value.
The online optimization stage operates on the basis of the following historical operating data: the method comprises the following steps of (1) updating relevant parameters of the remaining four control levels on line by a Particle Swarm Optimization (PSO) algorithm, wherein the parameters comprise the concentration of NOx at an SCR inlet of a boiler A/B side, the concentration of NOx at an SCR outlet of the boiler A/B side, the concentration of net smoke NOx, unit load, coal quantity, air quantity, ammonia flow and ammonia escape quantity, and the PSO algorithm is described in the following mathematic:
assuming a population size of N, the coordinate position of a particle i (i ═ 1,2, L, N) in a d-dimensional space can be represented as xi＝(xi1,xi2,L,xid) The velocity of the particle is defined as the distance the particle moves during each iteration, vi＝(vi1,vi2,L,vid) And (4) showing. At the k-th iteration, the flight velocity v of the particle i in the d-dimensional subspaceidThe adjustment update is performed according to the following formula:
the particles adjust their position by:
wherein the variables x respectively represent the relevant online optimization parameters of the four control levels;
for the fast protection level, the online optimization parameter x is a protection action amplitude coefficient;
for the intelligent feedforward stage, the online optimization parameters x are respectively corresponding linear variable parameter model input weight coefficientsAndweight biasAnda scaling factor δ;
for the precise control level, the online optimization parameters x are respectively the input weight coefficients of the corresponding linear variable parameter modelsAndweight biasAnda scaling coefficient delta and a proportional coefficient and an integral coefficient of the PID controller;
for the ammonia escape control level, the online optimization parameter x is a corrected value of the concentration of NOx at the outlet of the SCR at the A/B side of the boiler, and the initial value is + 3.
PSO algorithm flow chart, as shown in FIG. 2.
The hardware part of the online optimization control system is connected with a computer of a denitration system of a power plant by adopting a PLC (programmable logic controller), and good technical effects are generated through actual operation, and the implementation effect of the 660MW unit of the system is as follows:
1) short term precision control
The unit load is increased to 470MW to 516MW, the concentration setting value of the NOx at the SCR outlet of the A/B side of the boiler is changed to 41-35-39-43, and the maximum dynamic deviation<5mg/Nm3Deviation from steady state<2mg/Nm3The adjusting time is 5-8 min.
2) Ammonia slip control
The ammonia escape control level automatic intervention adjusting system automatically adjusts the correction value according to the ammonia escape mean value, and reduces the ammonia escape by 55.6% in the range of small influence of NOx at the SCR outlet of the A/B side of the boiler.
3) Medium term stability control
The 8h running condition is that the mean value of the concentration of the NOx in the clean flue gas is 40mg/Nm3(37-46), and NOx (292-470) mg/Nm at the inlet of SCR at the A/B side of the boiler3Maximum rate of change 20mg/Nm3And/min, the load changes 330-484 MW.
4) Medium and long term stability control
The 7-day running condition is that the mean value of the concentration of the NOx in the clean flue gas is 41mg/Nm3(35-47), boiler A/B side SCR inlet NOx (290-610) mg/Nm3And the load of the unit changes 330-484 MW, and the instantaneous and average values exceed the standard for 0 time.
5) Long term stable control
The 14-day running condition is that the mean value of the concentration of the NOx in the clean flue gas is 41mg/Nm3NOx (270-650) mg/Nm at inlet of SCR at A/B side of boiler3And the load of the unit changes by 280-620 MW, and the instantaneous and average values exceed the standard for 0 time.
Claims (6)
1. A multi-stage intelligent denitration online optimization control system based on data driving is characterized by comprising five control stages, namely a rapid protection stage, an intelligent feedforward stage, an accurate control stage, an ammonia escape control stage and an online optimization stage, wherein an ammonia flow deviation value of the total ammonia demand and the actual ammonia flow is obtained through the five control stages, and the opening degree of an ammonia flow regulating valve is correspondingly regulated according to the ammonia flow deviation value, so that the ammonia escape is controlled at the lowest value on the premise that the NOx emission does not exceed the standard;
wherein:
fast protection level: calculating ammonia demand through the concentration value and the change rate of the clean flue gas NOx to obtain primary ammonia demand, and rapidly increasing ammonia to prevent the over-standard phenomenon caused by sudden rising of the concentration of the NOx so as to control the NOx emission not to exceed the standard;
intelligent feed-forward stage: the method comprises the steps of outputting secondary ammonia demand through a feedforward control module formed by a linear variable parameter model structure by taking NOx concentration at an SCR inlet of a boiler A/B side, unit load, coal quantity, air quantity, measured value of SCR outlet concentration of the boiler A/B side and NOx concentration setting guidance value of the SCR outlet of the boiler A/B side as input variables, and directly changing ammonia flow when any variable changes so as to realize the rapidity of NOx emission control;
accurate control level: a set value prediction control loop and a cascade negative feedback prediction control loop;
the set value prediction control loop comprises a NOx concentration calculation module formed by a linear variable parameter model structure, a net flue gas NOx concentration set value is used as an input variable, the output is a boiler A/B side SCR outlet NOx concentration setting guide value, the boiler A/B side SCR outlet NOx concentration setting guide value is used as one variable of a feedforward control module in an intelligent feedforward level, the concentration setting guide value and a boiler A/B side SCR outlet NOx concentration correction value output by an ammonia escape control level are combined to obtain a boiler A/B side SCR outlet NOx concentration set value, the concentration set value is compared with a boiler A/B side SCR outlet NOx concentration measurement value to obtain a deviation value of the boiler A/B side SCR outlet NOx concentration set value and the measurement value, and the deviation value is used as an input quantity of a cascade negative feedback prediction control loop;
the cascade negative feedback prediction control loop comprises a prediction controller based on a compact generalized prediction model, the input of the prediction controller is the deviation value of a NOx concentration set value and a measured value at the outlet of an SCR (selective catalytic reduction) at the A/B side of a boiler, the output of the prediction controller is ammonia gas three-level demand, the nitrogen oxide value at the outlet of a denitration reactor under the current ammonia injection amount can be obtained by calculating in advance through the prediction model, whether the ammonia injection amount is proper or not is judged in advance, and then an adjustment suggested value of the ammonia injection amount is given;
summing the primary ammonia demand, the secondary ammonia demand and the tertiary ammonia demand to obtain the total ammonia demand;
transmitting the ammonia flow deviation value of the total ammonia demand and the actual ammonia flow to a PID (proportion integration differentiation) controller of a denitration system of the power plant for opening degree regulation control of an ammonia flow regulating valve;
ammonia slip control stage: correcting the NOx concentration setting guide value at the SCR outlet of the A/B side of the boiler according to the ammonia escape measured value to obtain the corrected value of the NOx concentration at the SCR outlet of the A/B side of the boiler, and improving the NOx emission concentration value at the SCR outlet to achieve the purpose of reducing ammonia escape;
and (3) online optimization stage: and updating the parameters of the rest four control levels on line through a particle swarm optimization algorithm according to historical operating data.
2. The multi-stage intelligent denitration online optimization control system based on data driving of claim 1, wherein the starting conditions of the rapid protection stage are as follows:
a. net flue gas NOx concentration value >47mg/Nm 3; b. the change rate of the net smoke NOx concentration value is more than 5mg/Nm 3/min;
when the two conditions of a and b are met simultaneously, the first-level demand of ammonia gas is output, and the output value is as follows:
ammonia gas current flow value x (net flue gas NOx current concentration value-47) x 5% x protection action amplitude coefficient
The over-standard phenomenon caused by the sudden rise of the concentration of the NOx can be prevented, and the NOx emission is controlled not to be over-standard.
3. The online optimization control system for multi-stage intelligent denitration based on data driving of claim 1 or 2, wherein the priority of the fast protection level is highest.
4. The data-driven multistage intelligent denitration online optimization control system according to claim 1, wherein the feedforward control module of the intelligent feedforward stage has a specific structure:
assuming that the input of the system at time t is u (t), the output is y (t), I and J are the orders of the model autoregressive part and the moving average part respectively, and ξ (t) is white noise, the linear parameter change model of the system can be expressed as:
wherein p, q is 1,2, …, n; 1,2, … I1；r＝1,2,…,m；j＝1,2,…,Jr(ii) a S is 1,2, …, S; m is the number of input variables, n is the number of output variables, and S is the number of selected reference working conditions;andin order to input the weight coefficient,is an exponential factor polynomial and is used as a reference,andin order to weight the offset, the weight is,is a function ofThe central point of (2), namely the selected reference working condition point; delta is a scaling factor, delta > 0.
In the control system, the number m of input variables is 6, the input variables are respectively the concentration u (1) of the NOx at the inlet of the SCR at the A/B side of the boiler, the unit load u (2), the coal quantity u (3), the air quantity u (4), the measured value u (5) of the concentration at the outlet of the SCR at the A/B side of the boiler and the set guidance value u (6) of the concentration at the outlet of the SCR at the A/B side, the number n of output variables is 1, the output y (1) is the secondary demand of ammonia gas, the reference working condition number S is 7, x is the time-varying parameter of the system, and the unit load is selected as the time-.
5. The online optimization control system for multistage intelligent denitration based on data driving of claim 1, wherein in the cascade negative feedback prediction control loop, the structure of the prediction controller based on the compact generalized prediction model is as follows:
y(k+j)＝Lj(q-1)Δu(k-1)+y0(k+j)
wherein: j ═ 1, Λ, P; p is the predicted step number; p is the predicted step number; k is the current time; l is the number of items in the control domain; u is the system input; y0 is the contribution of past inputs and outputs to future time output
The nitrogen oxide value at the outlet of the denitration reactor under the current ammonia injection amount can be calculated in advance through a prediction model.
6. The online optimization control system for multi-stage intelligent denitration based on data driving of claim 1, wherein the starting conditions of the ammonia escape control stage are as follows:
a. net flue gas NOx concentration<40mg/Nm3；
b. The set value of the ammonia injection valve is not changed within 10 min;
c. the average value of ammonia escape is more than 10 ppm;
when the three conditions a, B and c are met simultaneously, overlapping the corrected value of the concentration of the NOx at the SCR outlet of the A/B side of the boiler into the set guidance value of the concentration of the NOx at the SCR outlet of the A/B side of the boiler every 15min to achieve the purpose of improving the concentration value of the NOx, thereby reducing the ammonia escape;
after the ammonia escape control level intervenes in automatic control, if the net flue gas NOx concentration value is more than 45mg/Nm in 10min and the average value is more than 45mg/Nm3The correction values are not superimposed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910921188.6A CN110618706B (en) | 2019-09-27 | 2019-09-27 | Multistage intelligent denitration on-line optimization control system based on data driving |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910921188.6A CN110618706B (en) | 2019-09-27 | 2019-09-27 | Multistage intelligent denitration on-line optimization control system based on data driving |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110618706A true CN110618706A (en) | 2019-12-27 |
| CN110618706B CN110618706B (en) | 2023-05-12 |
Family
ID=68924295
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910921188.6A Active CN110618706B (en) | 2019-09-27 | 2019-09-27 | Multistage intelligent denitration on-line optimization control system based on data driving |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110618706B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112327958A (en) * | 2020-10-26 | 2021-02-05 | 江南大学 | Fermentation process pH value control method based on data driving |
| CN113304609A (en) * | 2021-05-28 | 2021-08-27 | 上海明华电力科技有限公司 | Balance control method for thermal power generating unit denitration system |
| CN114326387A (en) * | 2021-12-07 | 2022-04-12 | 江苏方天电力技术有限公司 | Denitration control device and method for thermal power generating unit |
| CN114534466A (en) * | 2020-11-11 | 2022-05-27 | 台泥资讯股份有限公司 | Method for controlling nitrogen oxide emission by control system |
| US11853023B2 (en) | 2020-11-11 | 2023-12-26 | Tcc Information Systems Corp. | Method of controlling coal management system for reducing coal usage |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH07227521A (en) * | 1994-02-16 | 1995-08-29 | Tokyo Gas Co Ltd | Method for preventing deposit of ammonium sulfate in selective reduction denitrification method |
| US20070122317A1 (en) * | 2005-11-30 | 2007-05-31 | Driscoll James J | Multi-stage system for selective catalytic reduction |
| US7824636B1 (en) * | 2009-07-24 | 2010-11-02 | General Electric Company | Model-based tuning of ammonia distribution and control for reduced operating cost of selective catalytic reduction |
| CN202087235U (en) * | 2011-03-28 | 2011-12-28 | 浙江大学 | Multi-level thermolysis coupling denitrification device with pre-flow field equalizing device |
| US20150052878A1 (en) * | 2013-08-20 | 2015-02-26 | Baohua Qi | Multi-stage SCR System |
| CN104785105A (en) * | 2015-04-07 | 2015-07-22 | 中国矿业大学 | Small-scale SCR (selective catalytic reduction) flue gas denitration ammonia injection control system and method |
| CN105629736A (en) * | 2016-03-22 | 2016-06-01 | 东南大学 | Data-driven thermal power generation unit SCR denitration disturbance suppression prediction control method |
| CN106582285A (en) * | 2016-12-29 | 2017-04-26 | 华电电力科学研究院 | Multi-stage ammonia spraying efficient SCR flue gas denitrification system |
| CN106621804A (en) * | 2016-11-15 | 2017-05-10 | 国家电网公司 | Fume ammonia injection mixing device and SCR (selective catalytic reduction) fume denitrification system |
| CN109062053A (en) * | 2018-08-31 | 2018-12-21 | 江苏国信靖江发电有限公司 | A kind of denitration spray ammonia control method based on multivariate calibration |
| CN109092035A (en) * | 2018-08-31 | 2018-12-28 | 华电电力科学研究院有限公司 | A kind of device and its working method for adjusting denitration spray ammonia flow adding feedforward based on cascade PID |
| CN209020167U (en) * | 2018-08-31 | 2019-06-25 | 华电电力科学研究院有限公司 | A kind of device for adjusting denitration spray ammonia flow adding feedforward based on cascade PID |
-
2019
- 2019-09-27 CN CN201910921188.6A patent/CN110618706B/en active Active
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH07227521A (en) * | 1994-02-16 | 1995-08-29 | Tokyo Gas Co Ltd | Method for preventing deposit of ammonium sulfate in selective reduction denitrification method |
| US20070122317A1 (en) * | 2005-11-30 | 2007-05-31 | Driscoll James J | Multi-stage system for selective catalytic reduction |
| US7824636B1 (en) * | 2009-07-24 | 2010-11-02 | General Electric Company | Model-based tuning of ammonia distribution and control for reduced operating cost of selective catalytic reduction |
| CN202087235U (en) * | 2011-03-28 | 2011-12-28 | 浙江大学 | Multi-level thermolysis coupling denitrification device with pre-flow field equalizing device |
| US20150052878A1 (en) * | 2013-08-20 | 2015-02-26 | Baohua Qi | Multi-stage SCR System |
| CN104785105A (en) * | 2015-04-07 | 2015-07-22 | 中国矿业大学 | Small-scale SCR (selective catalytic reduction) flue gas denitration ammonia injection control system and method |
| CN105629736A (en) * | 2016-03-22 | 2016-06-01 | 东南大学 | Data-driven thermal power generation unit SCR denitration disturbance suppression prediction control method |
| CN106621804A (en) * | 2016-11-15 | 2017-05-10 | 国家电网公司 | Fume ammonia injection mixing device and SCR (selective catalytic reduction) fume denitrification system |
| CN106582285A (en) * | 2016-12-29 | 2017-04-26 | 华电电力科学研究院 | Multi-stage ammonia spraying efficient SCR flue gas denitrification system |
| CN109062053A (en) * | 2018-08-31 | 2018-12-21 | 江苏国信靖江发电有限公司 | A kind of denitration spray ammonia control method based on multivariate calibration |
| CN109092035A (en) * | 2018-08-31 | 2018-12-28 | 华电电力科学研究院有限公司 | A kind of device and its working method for adjusting denitration spray ammonia flow adding feedforward based on cascade PID |
| CN209020167U (en) * | 2018-08-31 | 2019-06-25 | 华电电力科学研究院有限公司 | A kind of device for adjusting denitration spray ammonia flow adding feedforward based on cascade PID |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112327958A (en) * | 2020-10-26 | 2021-02-05 | 江南大学 | Fermentation process pH value control method based on data driving |
| CN114534466A (en) * | 2020-11-11 | 2022-05-27 | 台泥资讯股份有限公司 | Method for controlling nitrogen oxide emission by control system |
| US11853023B2 (en) | 2020-11-11 | 2023-12-26 | Tcc Information Systems Corp. | Method of controlling coal management system for reducing coal usage |
| CN114534466B (en) * | 2020-11-11 | 2024-03-22 | 台泥资讯股份有限公司 | Method for controlling nitrogen oxide emission by control system |
| CN113304609A (en) * | 2021-05-28 | 2021-08-27 | 上海明华电力科技有限公司 | Balance control method for thermal power generating unit denitration system |
| CN114326387A (en) * | 2021-12-07 | 2022-04-12 | 江苏方天电力技术有限公司 | Denitration control device and method for thermal power generating unit |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110618706B (en) | 2023-05-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110618706A (en) | Multistage intelligent denitration online optimization control system based on data driving | |
| US7813819B2 (en) | Control system for control subject having combustion unit and control system for plant having boiler | |
| CN106873381B (en) | ammonia spraying control system | |
| US8554706B2 (en) | Power plant control device which uses a model, a learning signal, a correction signal, and a manipulation signal | |
| JP4876057B2 (en) | Plant control device and thermal power plant control device | |
| CN107561941A (en) | A kind of full working scope qualified discharge control method of fired power generating unit denitrating system | |
| JP4270218B2 (en) | Control device for control object having combustion device, and control device for plant having boiler | |
| US20120143382A1 (en) | Optimized integrated controls for oxy-fuel combustion power plant | |
| JP2008146371A (en) | Controller of boiler plant | |
| JP4665815B2 (en) | Plant control equipment | |
| Xakimovich et al. | Mathematical modeling of transient processes of the automatic control system of water level in the steam generator | |
| Zhao et al. | A nonlinear industrial model predictive controller using integrated PLS and neural net state-space model | |
| JP5155690B2 (en) | Coal-fired boiler gas concentration estimation device and gas concentration estimation method | |
| CN110263452A (en) | Flue gas Annual distribution characteristic analysis method, system and denitrating system in a kind of flue | |
| JP5117232B2 (en) | Control device for plant with boiler and control method for plant with boiler | |
| CN108488831B (en) | Boiler combustion control system and method | |
| CN109833773A (en) | A kind of NO_x Reduction by Effective ammonia flow accuracy control method | |
| JP2012057809A (en) | Operation control method of incinerator | |
| CN115145152A (en) | Boiler combustion and denitration process collaborative optimization control method | |
| JP3653599B2 (en) | Apparatus and method for controlling ammonia injection amount of flue gas denitration equipment | |
| KR100194446B1 (en) | Combustion control method and apparatus for waste incinerator | |
| JP2000257824A (en) | Method and device for controlling combustion facility | |
| CN114673982B (en) | Main steam temperature control system of thermal power generation boiler based on hybrid intelligent optimization algorithm | |
| CN116203830A (en) | Air flow uncontrolled type heating furnace temperature PID optimization control method | |
| JP2840184B2 (en) | Incinerator combustion control device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |