CN103324991B - The method and system of pulverized-coal fired boiler NOx emission predictive - Google Patents

Abstract

The invention discloses the method and system of a kind of pulverized-coal fired boiler NOx emission predictive, by actual pulverized-coal fired boiler modelling, and model is carried out stress and strain model, then use FLUENT software to simulate boiler combustion process, and solve the concentration of NOx, in order to understand environmental.Owing to being made without test, eliminate experimental rig spending, and compare that traditional test method is more flexible, accuracy is higher.

Description

The method and system of pulverized-coal fired boiler NOx emission predictive

Technical field

The present invention relates to boiler technical field of power generation, particularly relate to a kind of method of pulverized-coal fired boiler Nox emitted smoke with System.

Background technology

Coal in pulverized-coal fired boiler while supplying energy, can with dust, sulfur dioxide, oxynitrides, carbon dioxide, The modes such as lime-ash pollute environment.Along with the sharp increase of China's industrial coal amount, the problem of environmental pollution caused by coal combustion is the most day by day Seriously, problem of environmental pollution has become restraining factors of China's electric power industry development.Moreover, it also causes seriously Social effects of pollution problem, such as form acid rain.

Above-mentioned environmental pollution is related to pulverized-coal fired boiler ring based on the oxynitrides i.e. discharge of NOx, the measurement to NOx The improvement of environment pollution problem.Traditional method utilizes boiler of power plant to run the experience accumulated, and small-scale experimental rig Measured data, assess the discharge capacity of NOx.The defect of this method is that result accuracy is low, tests both expensive.

Summary of the invention

Based on above-mentioned situation, the present invention proposes the method and system of a kind of pulverized-coal fired boiler Nox emitted smoke, with to coal dust The NOx discharge of boiler is predicted.

A kind of method of pulverized-coal fired boiler NOx emission predictive, including step:

Set up the model of pulverized-coal fired boiler；

The model set up is carried out stress and strain model；

Use FLUENT software that the combustion process of pulverized-coal fired boiler is simulated, determine the forecast model of NOx emission, and adopt By the concentration of NOx in described forecast model calculating stove.

A kind of system of pulverized-coal fired boiler NOx emission predictive, including:

Model building module, for setting up the model of pulverized-coal fired boiler；

Stress and strain model module, for carrying out stress and strain model to the model set up；

NOx prediction module, for using FLUENT software to be simulated the combustion process of pulverized-coal fired boiler, determines that NOx arranges The forecast model put, and use described forecast model to calculate the concentration of NOx in stove.

The method and system of pulverized-coal fired boiler NOx emission predictive of the present invention, by actual pulverized-coal fired boiler modelling, and to model Carry out stress and strain model, then use FLUENT software to simulate boiler combustion process, and solve the concentration of NOx, in order to understand environment dirty Dye degree.Owing to being made without test, eliminate experimental rig spending, and it is more flexible, accurate to compare traditional test method Property is higher.

Accompanying drawing explanation

Fig. 1 is that coal powder fired boiler burner region divides schematic diagram；

Fig. 2 is that the face of coal powder fired boiler burner fuel and air intake carries out stress and strain model schematic diagram；

Fig. 3 is cross section, coal powder fired boiler burner top stress and strain model schematic diagram；

Fig. 4 is Numerical Simulation of Turbulent method schematic diagram；

Fig. 5 is the schematic flow sheet of pulverized-coal fired boiler NOx emission predictive method of the present invention；

Fig. 6 is the structural representation of pulverized-coal fired boiler NOx emission predictive system of the present invention.

Detailed description of the invention

The present invention is with certain boiler of power plant as research object, by its combustion process of numerical simulation and temperature field, component field (include CO, O₂) and the discharge of NOx, and with Thermal-state test data as foundation, verify the accuracy of numerical simulation, propose corresponding The method of attenuating NOx emission.Specifically do following work.

Analyze model and the method for FLUENT software numerical simulation in stove the most in detail.

2. the topmost factor of the success of numerical simulation is the division of grid, includes unstrctured grid by experimental analysis With the impact of the result of the multiple stress and strain model logarithm value such as structured grid simulation, finally determine and a kind of coincide best with reality A kind of stress and strain model is as the basis of numerical computations.

3. pair result of calculation carries out the post processing of necessity, uses NOx model, by adjusting corresponding parameter, carries out numerical value Calculate, obtain a kind of distribution of the preferable NOx coincideing with reality.

4. the temperature that contrast Thermal-state test obtains with numerical simulation, various gas component, the distribution of NOx, analyze generation poor Different reason, and later improvement direction is proposed.

5. propose the measure of concrete minimizing NOx emission for the concrete type of furnace, and analyze related causes.

The stress and strain model of boiler body

It is respectively adopted unstrctured grid herein, with structured grid, the flowing in boiler has been carried out numerical simulation, find: non- Although structured grid has mess generation convenient, calculate the advantages such as very fast, but residual error fluctuation be relatively big, be difficult to convergence, and result with Structured grid is compared bigger than normal, and maximum temperature is about 2000k, and in using structured grid to calculate gained stove, maximum temperature is 1870k, and according to hot report from the point of view of, hence it is evident that be that the latter more meets reality.Therefore only introduce the division of structured grid below.

The stress and strain model of boiler body has two crucial places, and one is burner region, and another is furnace arch, furnace nose part Region.The stress and strain model situation of burner region is the most first discussed.

Owing to combustor section is the place that burning, flowing etc. are concentrated most, it is therefore necessary to combustion zone is carried out individually Stress and strain model.It is individual, as shown in Figure 1 that burner region is divided into 17 by the present invention.First comprise fuel to four to enter with air The body of mouth carries out stress and strain model.Stress and strain model is characterized by: for ensureing the symmetry of flowing, the surface grids of burner inlet is necessary There is certain symmetry, the most first the face of burner fuel Yu air intake is carried out stress and strain model, the result of division such as Fig. 2 Shown in, the specification of division is: 190mm*150mm.Then 1st district carries out the division of surface grids, and the specification of division is: 200mm* 200mm.Then complete the body comprising fuel and air intake is carried out the division of grid by cooper method.Then to comprise combustion Expect that 8 individualities adjacent with the body of air intake carry out the division of grid, comprise the concrete steps that the upper surface Tri first to 2nd district Primitive method generates the surface grids of 200mm*200mm, and this, primarily to prevent pseudo-diffusion, then generates by cooper method Corresponding volume mesh, the body in 3rd district first generates the surface grids of 200mm*200mm, then uses cooper upper surface by submap method Method generates volume mesh, and the body in 4th district directly generates the volume mesh of 200mm.The last net to whole upper surface divides, such as Fig. 3 institute Show.

Other body cooper method of boiler generates the hexahedral mesh of 200mm～300mm respectively, so collectively generates About more than 550,000 grid.

The stress and strain model of furnace arch, furnace nose part

Owing to the flowing of furnace arch, furnace nose part is the most more complicated, and there is turning, so stress and strain model here is to result also Have a great impact, furnace arch, furnace nose part through different three kinds of stress and strain model (other parameter constants), be computed having drawn below Several conclusions:

1. the air-flow eddy flow of outlet is mainly had a great impact, if dividing not by the stress and strain model of furnace arch, furnace nose part Good, by the lowest for the temperature making outlet region greatly, CO is dense.A certain degree of unburned charcoal is caused to lose so that The temperature of prediction is on the low side, and also can affect the distribution of other component fields.

2. body furnace arch, furnace nose part divided is the most, and result of calculation is the poorest.

3. the Meshing Method that furnace arch, furnace nose part uses herein can not make all of region on the cross section of exhanst gas outlet Temperature all meet reality, i.e. there is sub-fraction regional gas stream and be not screwed into, but result of calculation is not caused the biggest Impact.

Boundary condition and other parameters are arranged

Entrance uses speed entry condition, outlet to use pressure export condition.

The simulation of stove combustion process is mainly by the intercoupling of gas-particle two-phase, coupling between combustion reaction and turbulent flow The calculating of the Closure equation groups such as conjunction, momentum, energy and quality transmission draws the distribution situation of in-furnace temperature, flowing and component. The Stochastic Separated Flow Model that wherein solid phase particles uses, is added to particle among calculating as separation, and particle has independence mutually Equation group calculate, it is considered to the interaction between two-phase;Pulverized coal particle thinks that particle diameter keeps constant in combustion, Along with the quality that carries out of burning is constantly reducing.Turbulence model uses the K-ε model having swirling flow correction, burning to calculate and uses simplification Probability density function (PDF) calculate the component score in course of reaction, calculating being thought, the reaction of fuel and oxidant is one Individual fast reaction process, what PDF calculated is the Hybrid analysis of fuel and oxidant, the present invention in simulations using coal as fixed carbon With the mixture of volatile matter, double Hybrid analysis simulation of employing, PDF calculates the stoichiometric ratio form used, it is believed that fuel Fully react with oxidising agent.What the precipitation of volatile matter and reaction were followed is double competitive models, and what the reaction of coke was followed is Power includes scattered model.Furnace heat transmission 90% comes from radiation heat transfer, the P1 model that radiation patterns uses, it is considered to the spoke of particle Penetrate heat exchange.Wall surface temperature is set to 773K, and setting is to round plus after revising the temperature difference according to saturation water temperature, revises the temperature difference and examines That has considered that general combustion case slagging scorification causes heats.Pulverized coal particle item uses reflection type with the effect of wall, it is believed that particle touches Reenter flow field after there is minute surface elastic collision after bumping against wall and participate in flowing (the motion step number of particle is set to 20000). Wall calculation of boundary conditions processes according to Standard law of wall.Air and pulverized coal particle entrance are respectively provided with, and air port is according to speed Speed mouth processes, and arranges value and is shown in Table 1, and particle is that plain cloth is arranged on First air entry position mutually, the speed of setting and temperature and Secondary wind speed is same.Coal particle size is given according to Rosin-Rammler distribution, and minimum grain size is 9.999999e-06m, and maximum particle diameter is 200e-06m, average grain diameter is 75e-06m.Mass flow is that 1.12516kg/m3 outlet is set to pressure export, pressure export Pressure be 300pa.Solving form all uses single order upstreame scheme, various coefficient of relaxation to be set to: pressure coefficient: 0.3；Dynamic Coefficient of discharge: 0.5；Energy coefficient: 0.5；P1 coefficient: 0.8；Intermediate species coefficient: 0.7；Other coefficients select default value.Residual error is removed Energy equation is set to P1 model equation: 1e-06;Being set to of other: 1e-03.

In NOx forecast model, selecting heating power type, fuel type NOx, [0] prediction of thermal NO x uses partial equilibrium Method, Turbulent Parameter uses mixture parametric method.Char N in fuel type NOx: fugitive constituent N=0.7:0.3, char N uses NH3 mould Type, fugitive constituent N uses mixed model, and Beta surface area uses default value.

The ignition temperature of coal dust is set to 1800k, and fuel temperature is 333k, and beginning oxidizing temperature is 573k, Hybrid analysis Method uses 20 kinds of intermediate species.The industry of fuel, elementary analysis and low heat valve are shown in Table 2.

The parameter of table 1 1# entrance is arranged

Table 2 coal data

Project	Symbol	Unit	Operating mode one
				As-received moisture content	Mar	%	18.21
As-received ash	Aar	%	12.71
				As-received fugitive constituent	Var	%	23.53
As-received fixed carbon	Fcar	%	45.55
				As-received phosphorus content	Car	%	54.87
As-received hydrogen content	Har	%	3.43
				As-received nitrogen content	Nar	%	0.8
As-received sulfur content	Sar	%	0.36
				As-received oxygen content	Oar	%	9.62
Net calorific value as received basis	Qnet,ar	KJ/Kg	20993

FLUNET solves the convergence in temperature field and judges

1. the balance of the mass flow of the mass flow of entrance and outlet side

The total mass flow rate of entrance is: 122.4kg/s

The mass flow of outlet is (outlet): 133.4kg/s

The difference ratio of mass flow: (133.4-122.4)/133.4=8%, it is contemplated that have fraction pulverized coal particle to fall into ash Funnel, the result of calculation in this error range is the most reliably.

The most each residual values converges to a certain degree, and no longer changes with the increase of iterations.What FLUENT solved changes Generation number is more than 7000 time, and when iterations is about 700, the residual values of separate equation all tends to self-sustained oscillation, Qi Zhonglian Continuous equation falls below 10^-3Hereinafter, three rate equations fall below 10 with k-epsilon equation^-4Hereinafter, energy and radiation equation Fall below 10^-4Hereinafter, convergence can substantially be regarded as.

FLUNET solves the convergence of NOx concentration and judges

Owing to the growing amount of NOx is the least, the model that FLUENT uses is first to calculate temperature field, calculates NOx concentration the most again , therefore it is the least to import and export mass-flow change.The convergence situation of residual error is the most only discussed, by NO, HCN, NH₃Residual error convergence Figure understands, with the increase of iterations, NO, HCN, NH₃Residual error levels off to constant, and three kinds of residual errors all fall below 10^-6Hereinafter, This result being compared for 1000 times with iteration, outlet NO only reduces about 5ppm, is considered as result of calculation and has restrained.

The selection of solver

FLUENT provides separate type and manifold type two class solver, and manifold type is divided into implicitly and explicitly two kinds.

Separate type solver (segregated solver) is to solve each equation sequentially, seriatim (about u, v, w, p Equation with T).After namely first solving an equation (such as the u equation of momentum) on total-grid, then solve another one equation (such as the v equation of momentum).Owing to governing equation is nonlinear, and each other be coupling, therefore, obtain convergence solution it Before, be through too much taking turns iteration.Each iteration of taking turns is made up of following steps:

(1) according to the result of current solution, all flow variables are updated.Just start if calculated, then come with initial value Update.

(2) solve u, v and w equation of momentum the most respectively, obtain velocity field.When calculating, pressure and unit interface Mass flow uses current given value.

(3) speed obtained because of (2nd) step is likely to be unsatisfactory for continuity equation, therefore, with continuity equation and linearizing The pressure correction equation of the equation of momentum one Poisson type of structure, then solves this pressure correction equation, obtains pressure field and speed The correction value of degree field.

(4) utilize newly obtained velocity field and pressure field, solve the control of other scalars (such as temperature, tubulence energy and component etc.) Equation processed.

(5) for comprising the simulation of discrete phase, when inside exists Coupling Between Phases, according to the trajectory calculation result of discrete phase Update the source item of continuous phase.

(6) check whether equation group restrains.If not restraining, returning to (1st) step, repeating.

Manifold type solver (coupled solver) is to solve continuity equation, the equation of momentum, energy equation and group simultaneously Divide the coupled wave equation of transport equation, then, then solve turbulent flow equiscalar equation seriatim.Owing to governing equation is nonlinear, And be coupling each other, therefore, before obtaining convergence solution, be through 2 iteration.Each iteration of taking turns is by following steps group Become:

(1) according to the result of current solution, all flow variables are updated.Just start if calculated, then come with initial value Update.

(2) solve the Couple Equations (rear two of continuity equation, the equation of momentum, energy equation and component transport equation simultaneously Individual equation optionally solves).

(3) turbulent flow, radiation equiscalar equation as required, are solved seriatim.Note, before solving, equation being used The relevant variable result being previously obtained updates.

(4) for comprising the simulation of discrete phase, when inside exists Coupling Between Phases, according to the trajectory calculation result of discrete phase Update the source item of continuous phase.

(5) check whether equation group restrains.If not restraining, returning to (1st) step, repeating.

Two kinds of solvers are suitable for from being pressed onto the compressible a wide range of flowing of high speed, but generally, work as meter Calculate at a high speed can baric flow when moving, manifold type solver has more advantage.Manifold type solver usually can quickly be restrained, but required Internal memory be about 1.5 to 2 times of separate type solver.

Additionally, the several physical models provided in separate type solver, do not have in manifold type solver.These things Reason model includes: fluid volume model (VOF), multiphase mixture model, Euler's mixed model, PDF combustion model, pre-mixing combustion Model, the appointment periodic stream movable model of mass flow, cyclical heat conduction model and shell conduction model etc..

Owing to the present invention will use PDF combustion model to simulate boiler combustion, and computer hardware configuration is limited, therefore selects to divide From formula solver.

The selection of algorithm

FLUENT provides three kinds of algorithms: SIMPLE algorithm, SIMPLEC algorithm and PISO algorithm.SIMPLEC and PISO Algorithm has convergence rate faster, reduces 30%～50% relative to the calculating time of SIMPLE algorithm.FLUENT user's manual Recommending, for transient problem, PISO algorithm has obvious advantage, and for steady state problem, optional SIMPLE serial algorithm. The present invention selects SIMPLE algorithm.

The basic thought of SIMPLE algorithm can be described as follows: for given pressure field (value that it can assume that, or Result obtained by last iterative computation), solve the equation of momentum of discrete form, draw velocity field, therefore, it is necessary to given Pressure field revised.The principle revised is: the speed field energy corresponding with revised pressure field meets this iteration layer Continuity equation on secondary.Principle accordingly, we are in the relation generation by the pressure of the discrete form defined of the equation of momentum Yu speed Enter the discrete form of continuity equation, thus obtain pressure correction equation, pressure correction equation draw pressure correction value.Then, According to revised pressure field, try to achieve new velocity field.Then check whether velocity field restrains.If not restraining, with revised Force value, as given pressure field, starts the calculating of next level.The most repeatedly, until obtaining the solution of convergence.

Turbulent stress equation model

The governing equation of turbulent flow is:

Continuity equation:

$\frac{\∂\mathrm{\ρ}}{\∂t}+\mathrm{div}\left(\mathrm{\ρu}\right)=0$

The equation of momentum (Navier-Stokes equation):

$\frac{\∂\left(\mathrm{\ρu}\right)}{\∂t}+\mathrm{div}\left(\mathrm{\ρu}\stackrel{\→}{u}\right)=\mathrm{div}\left(\mathrm{\μ}\mathrm{grad}u\right)-\frac{\∂p}{\∂x}+[-\frac{\∂\left(\mathrm{\ρ}\stackrel{\‾}{u^{\′2}}\right)}{\∂x}-\frac{\∂\left(\mathrm{\ρ}\stackrel{\‾}{u^{\′}{v}^{\′}}\right)}{\∂y}-\frac{\∂\left(\mathrm{\ρ}\stackrel{\‾}{u^{\′}{w}^{\′}}\right)}{\∂z}]+S_u$

$\frac{\∂\left(\mathrm{\ρv}\right)}{\∂t}+\mathrm{div}\left(\mathrm{\ρv}\stackrel{\→}{u}\right)=\mathrm{div}\left(\mathrm{\μ}\mathrm{grad}v\right)-\frac{\∂p}{\∂y}+[-\frac{\∂\left(\mathrm{\ρ}\stackrel{\‾}{u^{\′}{v}^{\′}}\right)}{\∂x}-\frac{\∂\left(\mathrm{\ρ}\stackrel{\‾}{v^{\′2}}\right)}{\∂y}-\frac{\∂\left(\mathrm{\ρ}\stackrel{\‾}{v^{\′}{w}^{\′}}\right)}{\∂z}]+S_v$

$\frac{\∂\left(\mathrm{\ρw}\right)}{\∂t}+\mathrm{div}\left(\mathrm{\ρw}\stackrel{\→}{u}\right)=\mathrm{div}\left(\mathrm{\μ}\mathrm{grad}w\right)-\frac{\∂p}{\∂z}+[-\frac{\∂\left(\mathrm{\ρ}\stackrel{\‾}{u^{\′}{w}^{\′}}\right)}{\∂x}-\frac{\∂\left(\mathrm{\ρ}\stackrel{\‾}{v^{\′}{w}^{\′}}\right)}{\∂y}-\frac{\∂\left(\mathrm{\ρ}\stackrel{\‾}{w^{\′2}}\right)}{\∂z}]+S_w$

The transport equation of its dependent variable:

$\frac{\∂\left(\mathrm{\ρ\φ}\right)}{\∂t}+\mathrm{div}\left(\mathrm{\ρ}\stackrel{\→}{u}\mathrm{\φ}\right)=\mathrm{div}\left(\mathrm{\Γ}\mathrm{grad}\mathrm{\φ}\right)-\frac{\∂p}{\∂y}+[-\frac{\∂\left(\mathrm{\ρ}\stackrel{\‾}{u^{\′}{\mathrm{\φ}}^{\′}}\right)}{\∂x}-\frac{\∂\left(\mathrm{\ρ}\stackrel{\‾}{v^{\′}{\mathrm{\φ}}^{\′}}\right)}{\∂y}-\frac{\∂\left(\mathrm{\ρ}\stackrel{\‾}{w^{\′}{\mathrm{\φ}}^{\′}}\right)}{\∂z}]+S$

Visible, many six stress (3 direct stress and 3 shearing stress) in equation group, it may be assumed that

For the processing mode of these six stress, the method having obtained various Numerical Simulation of Turbulent, as Shown in Fig. 4.

The present invention selects Realizable k-ε model to be simulated.During standard k-ε model pair, equal strain rate is king-sized Situation, it is possible to cause the direct stress born.And Realizable k-ε model Main change is:

Turbulent viscosity computing formula there occurs change, the content introducing with rotating and curvature is relevant:

ε equation there occurs great changes, and the item that produces in equation no longer includes the generation item G in k equation_k, so, existing Form preferably illustrate spectrum energy conversion.

Item second from the bottom in ε equation does not have any singularity, even if k value is the least or be zero, denominator also will not be zero. This has the biggest difference with standard k-ε model and RNG k-ε model.

Realizable k-ε model must be applied to various types of flow simulating by effective, including rotating uniformly Shear flow, include flowing in jet and the free-flowing of mixed flow, pipeline, boundary layer flow, and with the flowing separated Deng.

Radiation patterns

FLUENT can calculate radiation heat transfer problem with 5 kinds of models, this in 5 model be discrete heat exchange radiation mode respectively Type (DTRM), P-1 radiation patterns, Rosseland radiation patterns, surface emissivity model (S2S) and discrete coordinates (DO) radiation mode Type.

For the burning that problem is boiler of present invention research, therefore select P-1 radiation patterns.

P-1 radiation patterns, radiation environment evaluation is one and calculates relatively small diffusion equation, contains simultaneously in model Scattering effect, in the computational problem that the optical thicknesses such as burning are the biggest, the calculating effect of P-1 model is relatively good.

Component transport and Chemical Reaction Model

FLUENT provides four kinds of analogy methods:

(1) general finite Rate Models

(2) non-premixed combustion model

(3) pre-mixing combustion model

(4) partial pre-mix combustion model

The substantially policy of Model Selection is as follows:

(1) general finite rate pattern is mainly used in: chemical constituent mixing, the problem transporting and reacting；Wall or grain The problem (such as chemical vapour deposition) of sub-surface reaction；

(2) non-premixed combustion model is mainly used in: include the reaction system of turbulent diffusion flame, this system closeization Learning balance, oxide therein and fuel separately flow into region to be calculated with two or three runners；

(3) pre-mixing combustion model is mainly used in: conjunction reagent flow single, fully premixed；

(4) partial pre-mix combustion model is mainly used in: have the feelings of the premixed flame of the equivalent ratio of change in region Condition；

The present invention selects non-premixed combustion model, with Hybrid analysis-probability density function simulation gas phase turbulent combustion.

NOx generates model

NO_xGenerate type and have three kinds: heating power type, Quick-type, fuel type.Formation mechanism is in chapter 1 introduction.Based on this Invention research object is the numerical simulation of coal dust firing, therefore does not consider Quick-type NO_x(it is mainly generated by CH fuel combustion), but Consider NO_xFuel efficiency again should.Additionally, model does not consider N₂The generation of O.

Discrete scheme

Pressure:PRESTO!

Momentum:Second Order Upwind

Turbulence Kinetic Energy:Second Order Upwind

Turbulence Dissipation Rate:Second Order Upwind

NO:Second Order Upwind

HCN:Second Order Upwind

NH₃:Second Order Upwind

Energy:Second Order Upwind

Mean Mixture Fraction:Second Order Upwind

Mixture Fraction Variance:Second Order Upwind

Coal composition

Wind speed, wind rate and wind-warm syndrome

Table 3 wind speed, wind rate and wind-warm syndrome

Burner arrangement

If pulverized-coal fired boiler is tangentially-fired boiler, burner has four groups, is arranged in four diagonal angles.Often organize burning Device divides six layers, respectively two-layer First air, three layers of Secondary Air and one layer of tertiary air.Primary and secondary air interval is arranged, tertiary air is arranged Topmost, interval, each air port is identical, is 550mm.Angle between four groups of burner axial lines and boiler diagonal is 4 ° 7 ', the circle of contact counterclockwise of a diameter of 850mm it is centrally formed at burner hearth.

Additionally, First air spout is vertical-type shade burner, air port is concentrated phase to fire side, and back-fire side is light phase.

Parameters for numerical simulation calculates

1, coal elemental composition mass fraction (%):

C:81.26

H:5.45

O:11.27

N:1.27

S:0.75

2, air quality mark (%):

N2:76.7

O2:23.3

Fuel Temperature:504K

Oxid Temperature:504K

3, Table Parameters:

Number of Mean Mixture Fraction Points:31

Number of Mixture Fraction Variance Points:21

Maximum Number of Species:20

Number of Mean Enthalpy Points:41

Minimum Temperature:298

Table 4 spout boundary condition

Table 5 muzzle velocity component on each axle

Wall boundary condition: wall surface temperature is set to 684K, material is steel (steel), and particle is had reflex (reflect)。

In the present invention, coal dust sliding velocity is set to the 10% of air velocity, i.e. coal dust speed is the 90% of air velocity.? Grain particle diameter obeys rosin-rammler rule.Injection source parameter such as table 6:

Injection Type:group

Particle Type:combusting

Stochastic Tracking:Discrete Random Walk Model

Table 6 injection source parameter

1、Mixture:

Thermal Conductive:0.025W/m.k

Viscosity:2e-05kg/m.s

Absorption Coefficient:wsggm-cell-based

2、Combusting-particles:

Density:1560kg/m3

Cp:1200J/kg.K

Thermal Conductivity:0.0454W/m.k

Latent Heat:0

Vaporization Temperature:333K

Volatile Component Fraction(%):29.3

Binary Diffusivity:5e-4m2/s

Particle Emissivity:0.9

Particle Scattering Factor:0.6

Swelling Coefficient:2

Burnout Stoichiometric Ratio:2.67

Combustible Fraction(%):56.5

Devolatilization Model:Two-rate Devolatilization Model

Combustion Model:kinectic/diffusion-limited

Number of Continuous Phase Iterations per DPM Iteration:20

Max.Number of Steps:26000

Length Scale:50mm

Simultaneously under Options, select Particle Radiation Interaction.

Beta PDF Points:25

Volatile N Mass Fraction:0.0188

Char N Mass Fraction:0.00805

BET Surface Area:25000

Below by as follows for the solution of the present invention combing:

The method of the pulverized-coal fired boiler NOx emission predictive of the present invention, as it is shown in figure 5, include step:

Step S101, set up the model of pulverized-coal fired boiler；

Step S102, the model set up is carried out stress and strain model；

The combustion process of pulverized-coal fired boiler is simulated by step S103, employing FLUENT software, determines the prediction of NOx emission Model, and use described forecast model to calculate the concentration of NOx in stove.

As a preferred embodiment, use structured grid that the model of pulverized-coal fired boiler is carried out stress and strain model, and to coal Burner region and the furnace arch, furnace nose part of powder boiler individually divide.

As a preferred embodiment, this method can also include step:

The NOx concentration calculating gained is compared with the NOx concentration of hot test gained, adjusts described pre-according to comparative result Survey model；

According to the NOx concentration of calculating gained, determine the modification scheme of coal powder fired boiler burner.

As a preferred embodiment, use volatile matter model, coke combustion model, radiation heat-transfer model, particle fortune Dynamic and gas phase flowing, gas phase turbulance combustion model are simulated the flowing in pulverized-coal fired boiler, and are obtained temperature field and component field distribution.

As a preferred embodiment, in described forecast model, select heating power type, fuel type NOx, thermal NO x [0] prediction use partial equilibrium method, Turbulent Parameter use mixture parametric method.Char N in fuel type NOx: fugitive constituent N= 0.7:0.3, char N uses NH₃Model, fugitive constituent N uses mixed model, and Beta surface area uses default value.

The system of pulverized-coal fired boiler NOx emission predictive of the present invention is corresponding with said method, as shown in Figure 6, and including:

Model building module, for setting up the model of pulverized-coal fired boiler；

Stress and strain model module, for carrying out stress and strain model to the model set up；

NOx prediction module, for using FLUENT software to be simulated the combustion process of pulverized-coal fired boiler, determines that NOx arranges The forecast model put, and use described forecast model to calculate the concentration of NOx in stove.

As a preferred embodiment, described stress and strain model module uses structured grid to carry out the model of pulverized-coal fired boiler Stress and strain model, and the burner region and furnace arch, furnace nose part to pulverized-coal fired boiler individually divide.

As a preferred embodiment, native system also includes:

Forecast model adjusting module, for the NOx concentration calculating gained is compared with the NOx concentration of hot test gained, Described forecast model is adjusted according to comparative result；

Burner reconstruction module, for the NOx concentration according to calculating gained, determines the transformation side of coal powder fired boiler burner Case.

As a preferred embodiment, described NOx prediction module uses volatile matter model, coke combustion model, radiation The motion of heat transfer model, particle and gas phase flowing, gas phase turbulance combustion model are simulated the flowing in pulverized-coal fired boiler, and are obtained temperature field With component field distribution.

As a preferred embodiment, described NOx prediction module, select heating power type, fuel type NOx, thermal NO x's [0] prediction uses partial equilibrium method, and Turbulent Parameter uses mixture parametric method.Char N in fuel type NOx: fugitive constituent N=0.7: 0.3, char N uses NH₃Model, fugitive constituent N uses mixed model, and Beta surface area uses default value.

Embodiment described above only have expressed the several embodiments of the present invention, and it describes more concrete and detailed, but also Therefore the restriction to the scope of the claims of the present invention can not be interpreted as.It should be pointed out that, for those of ordinary skill in the art For, without departing from the inventive concept of the premise, it is also possible to make some deformation and improvement, these broadly fall into the guarantor of the present invention Protect scope.Therefore, the protection domain of patent of the present invention should be as the criterion with claims.

Claims (10)

1. the method for a pulverized-coal fired boiler NOx emission predictive, it is characterised in that include step:

Set up the model of pulverized-coal fired boiler；

The model set up is carried out stress and strain model, specifically includes: burner region is divided into several body, to comprising fuel Carry out stress and strain model with the body of air intake, obtain the first division result of symmetry；After obtaining the first division result, to Comprise the fuel body adjacent with the body of air intake and carry out the division of grid, obtain the second division result；Two phases to next-door neighbour The adjacent body between body and a limit of burner carries out stress and strain model, obtains the 3rd division result；To burner remainder Body carry out stress and strain model, obtain the 4th division result；According to the first division result, the second division result, the 3rd division result With the 4th division result, the net obtaining whole upper surface divides；

Use FLUENT software that the combustion process of pulverized-coal fired boiler is simulated, determine the forecast model of NOx emission, and use institute State forecast model and calculate the concentration of NOx in stove.

The method of pulverized-coal fired boiler NOx emission predictive the most according to claim 1, it is characterised in that

Use structured grid that the model of pulverized-coal fired boiler carries out stress and strain model, and to the burner region of pulverized-coal fired boiler and furnace arch, furnace nose Part individually divides.

The method of pulverized-coal fired boiler NOx emission predictive the most according to claim 1 and 2, it is characterised in that further comprise the steps of:

The NOx concentration calculating gained is compared with the NOx concentration of hot test gained, adjusts described prediction mould according to comparative result Type；

According to the NOx concentration of calculating gained, determine the modification scheme of coal powder fired boiler burner.

The method of pulverized-coal fired boiler NOx emission predictive the most according to claim 1 and 2, it is characterised in that use volatile matter mould The motion of type, coke combustion model, radiation heat-transfer model, particle and gas phase flowing, gas phase turbulance combustion model simulation pulverized-coal fired boiler Interior flowing, and obtain temperature field and component field distribution.

The method of pulverized-coal fired boiler NOx emission predictive the most according to claim 1 and 2, it is characterised in that at described prediction mould In type, selecting heating power type, fuel type NOx, [0] prediction of thermal NO x uses partial equilibrium method, and Turbulent Parameter uses mixture Parametric method；Char N in fuel type NOx: fugitive constituent N=0.7:0.3, char N uses NH₃Model, fugitive constituent N uses hybrid guided mode Type, Beta surface area uses default value.

6. the system of a pulverized-coal fired boiler NOx emission predictive, it is characterised in that including:

Model building module, for setting up the model of pulverized-coal fired boiler；

Stress and strain model module, for the model set up is carried out stress and strain model, specifically includes: if being divided into by burner region Dry individuality, carries out stress and strain model to comprising the fuel body with air intake, obtains the first division result of symmetry；Obtaining first After division result, the body adjacent with the body of air intake with comprising fuel is carried out the division of grid, obtain the second division knot Really；Body between two adjacent bodies and a limit of burner of next-door neighbour is carried out stress and strain model, obtains the 3rd division result； The body of burner remainder is carried out stress and strain model, obtains the 4th division result；According to the first division result, the second division knot Really, the 3rd division result and the 4th division result, the net obtaining whole upper surface divides；

NOx prediction module, for using FLUENT software to be simulated the combustion process of pulverized-coal fired boiler, determines NOx emission Forecast model, and use described forecast model to calculate the concentration of NOx in stove.

The system of pulverized-coal fired boiler NOx emission predictive the most according to claim 6, it is characterised in that described stress and strain model mould Block uses structured grid that the model of pulverized-coal fired boiler carries out stress and strain model, and to the burner region of pulverized-coal fired boiler and furnace arch, furnace nose portion Divide and individually divide.

8. according to the system of the pulverized-coal fired boiler NOx emission predictive described in claim 6 or 7, it is characterised in that also include:

Forecast model adjusting module, for the NOx concentration calculating gained is compared with the NOx concentration of hot test gained, according to Comparative result adjusts described forecast model；

Burner reconstruction module, for the NOx concentration according to calculating gained, determines the modification scheme of coal powder fired boiler burner.

9. according to the system of the pulverized-coal fired boiler NOx emission predictive described in claim 6 or 7, it is characterised in that described NOx predicts Module uses the motion of volatile matter model, coke combustion model, radiation heat-transfer model, particle and gas phase flowing, gas phase turbulance burning Flowing in modeling pulverized-coal fired boiler, and obtain temperature field and component field distribution.

10. according to the system of the pulverized-coal fired boiler NOx emission predictive described in claim 6 or 7, it is characterised in that described NOx predicts Module, selects heating power type, fuel type NOx, and [0] prediction of thermal NO x uses partial equilibrium method, and Turbulent Parameter uses mixture Parametric method；Char N in fuel type NOx: fugitive constituent N=0.7:0.3, char N uses NH₃Model, fugitive constituent N uses hybrid guided mode Type, Beta surface area uses default value.