Intelligent prediction method and system for alkali metal slagging tendency of biomass gas coupled coal-fired boiler

Abstract

The application discloses a biomass gas coupling coal-fired boiler alkali metal slagging tendency intelligent prediction method and system, relates to the field of slagging tendency prediction, obtains flue gas temperature data of a target area of a boiler convection heating surface, obtains ash fusion characteristic temperature and alkali metal element content data of mixed fuel fed into the boiler, performs self-adaptive correlation calculation of ash temperature-liquid phase yield under physical constraints, calculates coupling adhesion force, generates particle surface equivalent adhesion strength distribution, based on the particle surface equivalent adhesion strength distribution, counts particle turbulent collision physical parameters, determines effective collision adhesion efficiency based on energy competition, calculates the agglomeration growth rate of ash particles in a flow field, outputs ash particle agglomerate spatial growth configuration, according to the ash particle agglomerate spatial growth configuration, performs macroscopic mechanical characteristic analysis and bridge stability determination on the accumulation body of the throat of a cold ash bucket, and outputs an ash bucket blockage risk level, thereby effectively solving the cold ash bucket blockage problem.

Description

Technical Field

[0001] This invention relates to the field of slagging tendency prediction, specifically to an intelligent prediction method and system for alkali metal slagging tendency in biomass gas coupled coal-fired boilers. Background Technology

[0002] In the actual operation of biomass gas coupled coal-fired power units, an atypical slag discharge anomaly, distinct from traditional high-temperature melting slag, is frequently observed in the convective heating surface and cold ash hopper area: although the flue gas temperature at the furnace outlet and in the convection zone is strictly controlled below the ash softening temperature and even the deformation temperature, theoretically molten coking should not occur, but severe structural bridging blockage frequently occurs at the throat of the cold ash hopper. On-site inspection revealed that this blockage is not the dense, dark, glassy hard slag commonly found in traditional coal-fired boilers, but rather a large, lightweight slag with extremely low macroscopic density. It is mostly yellowish-white in appearance, with a loose and porous cross-sectional structure, a dull sound when tapped, and exhibits significant flocculent agglomeration characteristics. This large lightweight slag is essentially composed of countless tiny fly ash particles in the flue gas flow field. Although they have not reached a thermodynamic state of overall melting and flow, they undergo physical capture and growth similar to a snowball effect through a certain low-temperature bonding mechanism, ultimately forming a huge porous aggregate at the ash hopper with structural strength sufficient to resist gravity collapse, causing the slag discharge machine to mechanically jam.

[0003] Existing technologies for predicting such risks often rely excessively on ash fusion characteristic temperatures determined based on the GB / T219 standard or conventional alkali-acid ratio indices. This results in significant scale bias and response lag in terms of physical mechanisms. The highly reactive alkali metal elements abundant in biomass fuels possess extremely strong mineral fluxing properties and readily undergo low-temperature eutectic reactions with aluminosilicates in coal ash. This leads to the precipitation of a micron-thick silicate liquid phase precursor film on the microscopically rough surface of fly ash particles even at temperatures far below the macroscopic deformation temperature (e.g., 800-900℃). While this trace liquid film is insufficient to alter the macroscopic morphology of the particles, it is enough to induce strong capillary forces and viscous dissipation forces at the moment of particle collision and contact, forming an irreversible low-temperature sintering neck. Existing prediction models often simplify fly ash into dry, smooth spheres, ignoring the key amplifying effect of microscopic rough texture on the particle surface on liquid film spreading and liquid bridge pinning. They cannot quantify this reaction-limited agglomeration process dominated by microscopic surface liquid phase viscosity, and therefore cannot accurately predict the risk of large light slag bridging in cold ash hoppers caused by low-temperature eutectic sintering.

[0004] To address the aforementioned shortcomings, a technical solution is provided. Summary of the Invention

[0005] To address the technical problems mentioned in the background section, this invention is proposed. This invention provides an intelligent prediction method and system for alkali metal slagging tendency in biomass gas-coupled coal-fired boilers.

[0006] The objective of this invention can be achieved through the following technical solutions:

[0007] A smart prediction method for alkali metal slagging tendency in biomass gas coupled coal-fired boilers includes:

[0008] Acquire flue gas temperature data for the target area of ​​the boiler's convective heating surface, and acquire ash melting characteristic temperature and alkali metal content data for the mixed fuels fed into the furnace.

[0009] Based on the flue gas temperature data, ash melting characteristic temperature and alkali metal element content data, a physically constrained adaptive correlation calculation of ash temperature-liquid phase yield is performed. A set of liquid bridge physical viscosity characteristic parameters is constructed by combining the micro-roughness characteristics of particle surface, the coupling adhesion resultant force is calculated, and the equivalent adhesion strength distribution of particle surface is generated.

[0010] Based on the equivalent adhesion intensity distribution on the particle surface, the physical parameters of particle turbulent collision are statistically analyzed, the effective collision adhesion efficiency based on energy competition is determined, the agglomeration growth rate of gray particles in the flow field is calculated, and the spatial growth configuration of gray particle agglomerates is output.

[0011] Based on the spatial growth configuration of the ash particle agglomerates, macroscopic mechanical characteristics analysis and bridging stability determination are performed on the accumulation in the throat of the cold ash hopper, and the ash hopper blockage risk level is output.

[0012] Furthermore, the step of generating the equivalent adhesion strength distribution on the particle surface is as follows:

[0013] Based on the flue gas temperature data, ash melting characteristic temperature and alkali metal element content data, a physically constrained adaptive correlation calculation of ash temperature-liquid phase yield is performed to determine the thickness of liquid phase precipitation on the particle surface.

[0014] Based on the three-dimensional morphology measurement data of particle surfaces, a residual height field is constructed by curvature stripping. The height difference structure function is calculated in the multi-scale effective analysis domain, and the equivalent roughness scaling index of particle surfaces is generated by log-linear regression analysis.

[0015] Furthermore, the step of generating the equivalent adhesion strength distribution on the particle surface also includes:

[0016] Based on the thickness of liquid phase precipitation on the particle surface, an equation for calculating the liquid bridge volume is constructed using the equivalent roughness scaling index of the particle surface and the surface tension coefficient of the liquid phase.

[0017] Based on the liquid bridge volume calculation equation, the effective wetting area and liquid bridge volume of the liquid phase in the contact area of ​​ash particles are calculated, and a set of liquid bridge physical viscosity characteristic parameters is generated.

[0018] Based on the set of physical viscosity characteristics parameters of the liquid bridge, an improved Rumpf particle adhesion theory model is used to calculate the resultant force of capillary force, viscous dissipation force and van der Waals force of the liquid bridge at the moment of particle contact, and generate the equivalent adhesion strength distribution on the particle surface.

[0019] Furthermore, the step of determining the thickness of the liquid phase precipitation on the particle surface is as follows:

[0020] The residence dose in the eutectic sensitive temperature zone is calculated using the flue gas temperature data, and the effective alkali metal index is obtained by performing an activity correction on the alkali metal element content data using the ash melting characteristic temperature.

[0021] The effective liquid volume fraction was calculated by adaptively correcting the preset reference gray temperature-liquid phase yield relationship using the residence dose and effective alkali metal index in the eutectic sensitive temperature zone as physical constraint factors.

[0022] Based on the particle geometric characteristic parameters, a conversion relationship between volume fraction and film thickness is constructed, and the effective liquid phase volume fraction is converted into the liquid phase precipitation thickness on the particle surface.

[0023] Furthermore, the step of generating the equivalent roughness scaling index of the particle surface includes:

[0024] Based on the height field data measured in three dimensions, a servo-driven solid coordinate system and a circular surface solid domain are established. Within the domain, macroscopic curvature is stripped away through low-order shape fitting and differential operations to generate a residual height field solid.

[0025] The residual height field entity is used to construct a multi-scale annular core domain. Geometric Boolean operations are performed with the projected measurement area and the mass mask to obtain the effective annular domain. The effective sector set is then obtained by filtering based on the arc length ratio threshold.

[0026] Furthermore, the step of generating the equivalent roughness scaling index of the particle surface also includes:

[0027] Within the set of effective sectors, perform homogeneous sampling constraints, calculate the height difference structure function that satisfies the point-to-point spacing constraints, and generate physical scales corresponding to each scale.

[0028] Based on the instrument's physical limits and boundary effects, a clipping threshold is set, and the effective scale segment of the physical scale is truncated. Then, linear regression fitting is performed in the logarithmic domain to output the equivalent roughness scaling index of the particle surface.

[0029] Furthermore, the step of outputting the spatial growth configuration of gray particle aggregates includes:

[0030] Based on the flue geometry of the boiler's convective heating surface and the average flue gas velocity, the collision parameters of ash particles are estimated using empirical formulas based on turbulence theory, and the particle collision frequency and relative impact velocity are statistically analyzed.

[0031] The particle collision kinetic energy is calculated based on the relative impact velocity. The particle collision kinetic energy is then compared with the critical adhesion energy, which is transformed from the equivalent adhesion intensity distribution on the particle surface, to calculate the effective collision adhesion efficiency.

[0032] Furthermore, the step of outputting the spatial growth configuration of gray particle aggregates also includes:

[0033] Based on particle collision frequency and effective collision adhesion efficiency, the characteristic aggregation rate integral method is used to calculate the degree of aggregation growth of gray particles in the flow field, and generate the spatial growth configuration of gray particle aggregates.

[0034] Furthermore, the step of determining the risk level of clogging in the output ash hopper includes:

[0035] Based on the spatial growth configuration of aggregates and combined with the thickness of liquid phase precipitation on particle surface, the physical properties of the accumulation body in the throat of cold ash hopper are estimated and analyzed in an engineering manner to obtain the macroscopic equivalent density and overall shear strength.

[0036] Based on macroscopic equivalent density and overall shear strength, a competitive analysis logic for bridge stability is constructed to determine the risk level of ash hopper blockage.

[0037] A smart prediction system for alkali metal slagging tendency in biomass gas coupled coal-fired boilers includes:

[0038] The data acquisition module is used to acquire flue gas temperature data of the target area of ​​the boiler's convective heating surface, and to acquire ash melting characteristic temperature and alkali metal content data of the mixed fuel entering the furnace.

[0039] The adhesion distribution module is used to perform physically constrained adaptive correlation calculation of ash temperature-liquid phase yield based on the flue gas temperature data, ash melting characteristic temperature and alkali metal element content data, construct a set of liquid bridge physical viscosity characteristic parameters in combination with particle surface micro-roughness characteristics, calculate the coupling adhesion resultant force, and generate the equivalent adhesion strength distribution of particle surface.

[0040] The aggregate space module is used to statistically analyze the particle turbulent collision physical parameters based on the equivalent adhesion intensity distribution on the particle surface, determine the effective collision adhesion efficiency based on energy competition, calculate the agglomeration growth rate of gray particles in the flow field, and output the spatial growth configuration of gray particle agglomerations.

[0041] The risk assessment module is used to perform macroscopic mechanical characteristic analysis and bridging stability assessment on the accumulation body at the throat of the cold ash hopper based on the spatial growth configuration of the ash particle agglomerates, and output the ash hopper blockage risk level.

[0042] Compared with the prior art, the beneficial effects of the present invention are:

[0043] This invention acquires flue gas temperature data of the target area of ​​the boiler's convective heating surface, and obtains ash melting characteristic temperature and alkali metal content data of the mixed fuel entering the furnace. Based on the flue gas temperature data, ash melting characteristic temperature, and alkali metal content data, it performs a physically constrained adaptive correlation calculation of ash temperature and liquid phase yield. Combining the micro-roughness characteristics of the particle surface, it constructs a set of physical viscosity parameters of the liquid bridge, calculates the coupling adhesion resultant force, generates the equivalent adhesion intensity distribution of the particle surface, and based on the equivalent adhesion intensity distribution of the particle surface, it statistically analyzes the particle turbulent collision physical parameters, determines the effective collision adhesion efficiency based on energy competition, calculates the agglomeration growth rate of ash particles in the flow field, and outputs the spatial growth configuration of ash particle agglomerates. Based on physically constrained adaptive correlation calculations of ash temperature and liquid phase yield, this invention accurately quantifies the thickness of the liquid phase precipitation on the particle surface of biomass ash induced by the high activity of alkali metals in the eutectic sensitive temperature range (800-900℃), i.e., the trace liquid phase film, directly locking in the microscopic material basis leading to low-temperature sintering. Utilizing an improved Rumpf particle adhesion theory model, this invention corrects the effective wetting area by using an equivalent roughness scaling index obtained from three-dimensional morphology measurements. This substantially incorporates the changes in liquid bridge volume distribution caused by micron-level roughness texture and the resulting liquid bridge pinning effect in mechanical calculations, achieving accurate calculation of the coupling adhesion force (including capillary force, viscous dissipation force, and van der Waals force) of fly ash particles at the moment of contact.

[0044] Based on the spatial growth configuration of ash particle agglomerates, this invention performs macroscopic mechanical characteristic analysis and bridging stability determination on the accumulation at the throat of the cold ash hopper, outputting the ash hopper blockage risk level. Utilizing Abrahamson's inertial collision theory and the confined agglomeration mechanism, the dynamic process of ash particles forming spatial growth configurations of ash particle agglomerates in the flow field is reproduced. Based on this, a dual porosity model incorporating micropores of the agglomerates and macroscopic voids in the bed is constructed, theoretically deriving the key physical property that the accumulation has extremely low macroscopic equivalent density. Based on the mechanical equilibrium model of Jenike's bulk material flow theory, this invention further constructs a dimensionless bridging tendency index. By comparing the overall shear strength with gravity load, the high-risk characteristics of large light slag bridging are intuitively quantified, thus effectively solving the problem of cold ash hopper blockage caused by the contradictory characteristics of extremely low density and low-temperature sintering, which is difficult to predict using traditional methods. Attached Figure Description

[0045] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. The following drawings are not drawn to scale according to the actual size, but are intended to illustrate the main idea of ​​the present invention.

[0046] Figure 1A flowchart of a method for intelligent prediction of alkali metal slagging tendency in biomass gas coupled with coal-fired boilers;

[0047] Figure 2 System block diagram of a biomass gas coupled coal-fired boiler alkali metal slagging tendency intelligent prediction system;

[0048] Figure 3 This is a schematic diagram illustrating the principle of separating the motional solid-state coordinate system from the curvature.

[0049] Figure 4 This is a schematic diagram of multi-scale ring-band Boolean operations and filtering logic. Detailed Implementation

[0050] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are also within the scope of protection of the present invention.

[0051] Example 1

[0052] like Figure 1 As shown, the intelligent prediction method for alkali metal slagging tendency in biomass gas coupled coal-fired boilers includes the following steps:

[0053] Step S100: Obtain flue gas temperature data of the target area of ​​the boiler convective heating surface, and obtain ash melting characteristic temperature and alkali metal element content data of the mixed fuel entering the furnace;

[0054] Specifically, based on boiler structural layout and operating experience, the locations in the boiler's convective heating surfaces prone to slagging and high-temperature corrosion are identified as target areas. These include, for example, the flue gas passage section between the furnace outlet and the first convective heating surface, or the front end of the economizer or superheater. Flue gas temperature measuring points are arranged at the inlet and outlet positions of this target area. These measuring points can employ thermocouples, fiber optic temperature probes, or other commonly used industrial online temperature measuring devices. Through the boiler's distributed control system or independent data acquisition system, temperature data along the flue gas flow direction within the target area under typical load conditions is continuously recorded to obtain a flue gas temperature data sequence reflecting the temperature distribution of this area. The arithmetic mean of the temperatures from multiple measuring points along the path is taken to obtain an average flue gas temperature representing the overall level of the target area. This average flue gas temperature is used as the flue gas temperature data for subsequent calculations. Sampling and laboratory analysis are performed on the mixed fuels fed into the furnace (e.g., a mixture of pulverized coal and biomass). The collected fuel samples are prepared into ash samples using a standard ash preparation process. The characteristic melting temperatures of these ash samples, including softening temperature, hemispherical temperature, and flow temperature, are determined according to current standard ash fusion characteristic test methods. These temperatures characterize the typical temperature features of the transition of the fuel ash from a solid phase to a highly fluid liquid phase. Conventional chemical analysis methods such as X-ray fluorescence analysis, inductively coupled plasma atomic emission spectrometry, and ion chromatography are used to determine the mass fractions of oxides such as K₂O and Na₂O in the ash samples. Alternatively, elemental analysis methods are used to determine the mass fractions of alkali metal elements such as potassium (K) and sodium (Na), and these are converted into the total alkali metal content per unit mass of ash. Through these steps, three types of data are output: representative flue gas temperature data for the target area of ​​the convective heating surface; characteristic melting temperature data for the mixed fuel ash; and alkali metal content data converted to an ash basis. These data will serve as the basis for calculating the resident dose and effective alkali metal index in subsequent step S2011.

[0055] The intelligent prediction method for alkali metal slagging tendency in biomass gas coupled coal-fired boilers also includes the following steps:

[0056] Step S200: Based on the flue gas temperature data, ash melting characteristic temperature and alkali metal element content data, perform physically constrained ash temperature-liquid phase yield adaptive correlation calculation, construct a set of liquid bridge physical viscosity characteristic parameters in combination with particle surface micro-roughness characteristics, calculate the coupling adhesion resultant force, and generate the equivalent adhesion strength distribution on the particle surface.

[0057] Step S200, which involves generating the equivalent adhesion strength distribution on the particle surface, specifically includes the following steps:

[0058] Step S201: Based on the flue gas temperature data, ash melting characteristic temperature and alkali metal element content data, perform a physically constrained adaptive correlation calculation of ash temperature-liquid phase yield to determine the thickness of liquid phase precipitation on the particle surface.

[0059] Performing physically constrained adaptive correlation calculations for gray temperature and liquid phase yield specifically includes the following steps:

[0060] Step S2011: Calculate the residence dose in the eutectic sensitive temperature zone using the flue gas temperature data, and perform an activity correction on the alkali metal element content data using the ash melting characteristic temperature to obtain the effective alkali metal index;

[0061] The effective liquid volume fraction was calculated by adaptively correcting the preset reference gray temperature-liquid phase yield relationship using the residence dose and effective alkali metal index in the eutectic sensitive temperature zone as physical constraint factors.

[0062] Specifically, based on the softening temperature, hemispherical temperature, and other melting characteristic temperatures of ash, as well as existing melting test experience, a temperature range for a eutectic sensitive temperature zone is determined. This temperature zone characterizes the temperature range within which alkali metals in the ash form low-melting-point eutectics with aluminosilicates such as SiO2 and Al2O3, readily generating a liquid phase. For example, a symmetrical temperature band near the softening temperature can be selected as the eutectic sensitive temperature zone, or the 800℃–900℃ range can be chosen as the preferred eutectic sensitive temperature zone based on experimental results. Combining the geometric dimensions of the target area of ​​the boiler's convective heating surface and the flue gas velocity under actual operating conditions, the time required for fly ash particles to pass through the eutectic sensitive temperature zone is estimated, i.e., the average residence time within this temperature zone. The geometric dimensions of the target area of ​​the convective heating surface include the length of the flue gas passage and the flow cross-sectional area. When the velocity does not vary significantly along the path, the average residence time is obtained by dividing the effective length by the average flow velocity. Multiplying the average residence time by the ash mass flow rate through the region per unit time yields the ash residence dose within the eutectic sensitive temperature zone. This residence dose expresses the combined intensity of thermal and mass migration effects on ash within a temperature zone conducive to eutectic formation under given operating conditions. The ash mass flow rate of the region refers to the total mass of fly ash particles accompanying flue gas flow through the target area of ​​the convective heating surface per unit time. Based on melting characteristic temperature and alkali metal content data, an activity correction is applied to the alkali metal content. Specifically, when the ash softening temperature and hemispherical temperature are significantly lower than those of conventional low-alkali coal ash, it indicates that the ash is more prone to softening and melting at lower temperatures. In this case, the same alkali metal content has a stronger effect on promoting liquid phase formation, and is assigned a higher activity weight accordingly. Conversely, when the melting characteristic temperature is higher, the alkali metal activity is considered relatively low, and a lower activity weight is assigned accordingly. By multiplying the alkali metal content converted to the ash baseline by the activity weight obtained based on the melting characteristic temperature, a single effective alkali metal index can be obtained. This index is used to comprehensively characterize the combined effect of alkali metal content and ash fusibility. This invention does not limit the specific mathematical expression. Those skilled in the art can construct the activity weight and effective alkali metal index using common forms such as linear, piecewise linear or exponential, as long as the value of the effective alkali metal index is larger when the melting characteristic temperature is lower and the alkali metal content is higher.

[0063] The residence dose in the eutectic sensitive temperature zone With effective alkali metal index Together, they act as physical constraint factors, influencing the pre-established baseline gray temperature-liquid phase yield relationship, which can be expressed as: ,in, This represents the basic liquid phase volume fraction per unit ash volume, without considering the influence of high-alkali metals. This indicates the temperature at which the ash content is located, i.e., ash temperature. This is an empirical function obtained by fitting experimental data of low-alkali metal standard ash samples, used to describe the ash temperature at which the metal is melted. The fundamental variation law of liquid phase volume fraction within the ash content; in this invention, utilizing... and By adaptively correcting the key parameters in this baseline relationship, an effective gray temperature-liquid phase yield relationship can be obtained. For example, a temperature correction coefficient can be defined. and amplitude correction factor The effective liquid volume fraction was obtained. Its calculation formula is ,in The effective liquid volume fraction is used to characterize the volume fraction of ash particles per unit volume that can be converted into liquid phase, after considering the influence of high-alkali metals and actual operating conditions. To determine the representative flue gas temperature within the target region of the convective heating surface or the eutectic sensitive temperature zone, the average temperature of that region, the central temperature of the temperature range, etc., can be selected or combined as needed. This is a temperature correction factor used to shift the effective temperature at which the liquid phase begins to appear in the baseline gray temperature-liquid phase yield relationship towards lower temperatures, or to change the initial range of liquid phase growth with increasing temperature, when the residence dose and effective alkali metal index increase in the eutectic sensitive temperature region. This is an amplitude correction factor used to increase the upper limit of liquid phase volume fraction at a given representative flue gas temperature when the resident dose and effective alkali metal index increase, so that the actual liquid phase formation level under high alkali conditions is higher than that under the baseline conditions; this invention does not limit the temperature correction factor. and amplitude correction factor The specific mathematical form of the correction coefficients can be constructed by those skilled in the art based on experimental data or engineering experience, using monotonic functions such as linear, piecewise linear, or exponential functions, provided that when... and When increased, the temperature correction factor With amplitude correction factor The changing trend should reasonably reflect the decrease in the liquid phase formation initiation temperature and the increase in liquid phase formation intensity. Through the above adaptive correction of coefficients, an effective ash temperature-liquid phase yield relationship reflecting the actual mixed fuel characteristics and actual operating conditions can be obtained from the original baseline ash temperature-liquid phase yield relationship. This will represent the flue gas temperature in the target area of ​​the convective heating surface. Substituting the corrected formula, the corresponding effective liquid volume fraction can be obtained. The result This will serve as an input parameter for subsequent steps, including liquid phase precipitation thickness conversion and slagging risk assessment. Based on this, the liquid phase generation level of high-alkali fuel under actual boiler operating conditions can be more accurately characterized.

[0064] Step S2012: Construct a conversion relationship between volume fraction and film thickness based on particle geometric characteristic parameters, and convert the effective liquid phase volume fraction into the liquid phase precipitation thickness on the particle surface.

[0065] Specifically, based on the particle size analysis results of the mixed fuel ash entering the furnace or boiler operating experience, typical fly ash particle geometric characteristic parameters that can represent the deposition behavior of the convective heating surface are selected. Assuming the fly ash particles are equivalent spherical particles, a representative average particle size is determined based on the particle size analysis results, and half of this average particle size is taken as the equivalent radius. For cases requiring more precise description, several representative particle size segments can be divided according to the particle size range, and the equivalent radius of each particle size segment can be determined separately. The aforementioned equivalent radius is used to approximately characterize the external surface area and volume of the particles. In the initial stage of ash melting, the liquid phase is mainly distributed on the outer surface of the fly ash particles and in the contact area between the particles. Furthermore, in the stage where the overall liquid phase content is relatively low and large-scale melting agglomeration has not yet formed, the liquid phase preferentially spreads along the particle surface into a continuous film, rather than accumulating in large quantities in the internal pores of the particles or as independent droplets far from the particle surface. Therefore, under typical boiler operating conditions, especially in the stage dominated by early deposition and initial adhesion, the total liquid phase volume per unit volume of the particle system is mainly contributed by the volume of the liquid phase film on the particle surface. Based on this, the present invention adopts a particle description model based on the assumption that the liquid phase is distributed in the form of a surface film, and regards the liquid phase characterized by the effective liquid phase volume fraction as a liquid phase film uniformly coated on the outer surface of the particles, in order to establish an engineering conversion relationship from volume fraction to film thickness.

[0066] Using an equivalent spherical particle model, the equivalent radius of the gray particle is denoted as . The surface area of ​​a single particle is approximately When the average thickness of the particle surface is When a liquid phase film is formed, the liquid volume on the surface of a single particle can be approximately expressed as the product of the surface area and the film thickness, i.e. In a unit volume particle system, let the number of particles be . The total volume of the liquid phase in the system is approximately Let the total volume of particulate ash in this system be . The present invention uses the effective liquid volume fraction The ratio of the total volume of the liquid phase to the total volume of the particulate ash is given by: According to the law of conservation of volume, Substituting into the above definition of volume fraction, and combining it with the equivalent spherical particle volume... With the number of particles The relationship By representing the average liquid phase precipitation thickness, we can obtain the average liquid phase precipitation thickness. With effective liquid volume fraction The approximate linear relationship between them can be simplified to only be expressed as... and Related formats: Therefore, the above engineering conversion formula is used to characterize the average liquid phase precipitation thickness on the particle surface, where The coefficient 1 / 3 is the thickness of liquid phase precipitation on the particle surface. It is determined by the ratio of the volume to the surface area of ​​the equivalent spherical particle and the liquid phase volume conservation condition. It can reasonably reflect the order of magnitude and trend of liquid phase precipitation thickness under typical working conditions.

[0067] Steps S2011 and S2012, through a cross-scale correlation from the macroscopic thermochemical environment to the microscopic particle surface physical state, precisely quantified the fundamental cause of low-temperature sintering—the trace liquid phase film. Specifically, step S2011, based on the chemical mechanism that alkali metal elements in biomass ash readily undergo slow eutectic reactions with aluminosilicates in the temperature range below the ash melting point (i.e., the eutectic sensitive temperature range of 800-900℃), introduces the "residual dose" to characterize the cumulative effect of reaction time and the "effective alkali metal index" to characterize component activity. These two physical constraint factors are used to adaptively correct the baseline ash temperature-liquid phase yield curve (i.e., shift the initial temperature and increase the liquid phase amplitude), thereby theoretically explaining and... The calculation determined why an effective liquid phase volume fraction sufficient to generate viscosity had been generated inside the ash particles under seemingly unmelted low-temperature conditions. Step S2012, based on the physical fact that the liquid phase preferentially spreads along the particle surface rather than agglomerates during the initial precipitation stage, constructed an equivalent spherical particle surface wetting model. Using the principle of volume conservation, an engineering conversion relationship from volume fraction to geometric film thickness was established, transforming the abstract liquid phase content into the specific liquid phase precipitation thickness on the particle surface. This thickness parameter directly determines the ability of liquid bridges to form and the magnitude of capillary pull during subsequent particle collisions, providing the most basic microscopic physical boundary conditions for revealing how countless tiny ash particles are bonded into a huge slag block with a loose structure but strong cohesion through the snowball effect.

[0068] Step S202: Based on the three-dimensional morphology measurement data of the particle surface, a residual height field is constructed by curvature stripping. The height difference structure function is calculated in the multi-scale effective analysis domain. Log-linear regression analysis is used to generate the equivalent roughness scaling index of the particle surface.

[0069] like Figure 3 The diagram shown illustrates the principle of establishing a motional solid-state coordinate system and curvature stripping.

[0070] Reference Figure 3 The solid blue line at the top of the diagram represents the original height H(x,y), indicating the original surface morphology of the particles obtained by the 3D morphology measurement device. The dashed orange line in the middle of the diagram represents the trend surface. (Macroscopic curvature) represents the macroscopic shape of the particle obtained through low-order shape fitting; the green line below the figure represents the residual height. (The flattened micro-texture) represents the pure micro-roughness and undulations retained after removing the macro-curvature.

[0071] Reference Figure 3 A kinematic solid-state coordinate system is established, and the analysis of particle surface microstructure is used as an example for description. Local reference point. The red dot in the diagram represents the selected local reference point. This point is located in a region on the particle surface where the position is stable and the curvature change is gradual, such as the point of minimum curvature gradient modulus. A coordinate system is established using this point as the origin. Kinematic solid-state coordinate system: As shown by the arrow in the diagram, the reference point... The local normal direction defines the local height axis w (normal), and the lateral coordinate axes u and v (tangential) are defined in a tangential plane perpendicular to w. This coordinate system moves with the local curvature of the particle surface to eliminate the influence of the particle's macroscopic attitude.

[0072] Curvature stripping and residual extraction fitting of macroscopic trend surface (Step 1): The system is based on Within the central circular patch entity domain, a low-order fit (such as quadratic surface fitting) is performed on the original height data to generate the trend surface shown by the orange dashed line in the figure. This trend surface reflects the overall spherical or ellipsoidal profile of the particles. Difference operation (Step 2): As shown by the downward arrow in the figure, the system performs a difference operation (w- This involves subtracting the trend surface height from the original height. Obtaining the microscopic residual field (Step 3): After the difference operation, the residual height shown below is obtained. This process is mathematically equivalent to virtually flattening the curved particle surface, stripping away low-frequency macroscopic curvature information and retaining only high-frequency microscopic texture features, thus providing a pure geometric benchmark for subsequent roughness analysis.

[0073] Step S2021: Based on the height field data measured in three dimensions, establish a follower solid coordinate system and a circular surface solid domain. Within the domain, remove the macroscopic curvature through low-order shape fitting and difference operations to generate a residual height field solid.

[0074] The surface of a single solid particle is scanned using white light interferometry, confocal microscopy, structured light 3D scanner, laser displacement sensor array, or other 3D topography measurement devices to obtain particle surface height field data expressed in laboratory global coordinates. The height field data includes a set of spatially discrete points on the particle surface and their corresponding height values. This indicates that the height field is a function of the global plane coordinates of the laboratory as the independent variable. The corresponding horizontal position in the laboratory coordinate plane The position corresponding to the vertical or another orthogonal direction in the laboratory coordinate plane. Represents coordinates in the global plane The measured height value at the particle surface. (Refer to...) Figure 3 A point on the particle surface that is stable in position and has a gentle change in curvature is selected as a local reference point. Among them, the calculation of height field data is in The geometric centroid of the planar projection region is used as the center to define a search neighborhood. For example, the radius of this search neighborhood is 10% of the equivalent radius of the particle. This neighborhood ensures that the selected point is far from the noise region at the particle edge, thus possessing positional stability. The local average curvature or Gaussian curvature of each data point within the search neighborhood is calculated, and the local gradient modulus of the curvature is further calculated. The point with the smallest local gradient modulus of curvature within the neighborhood is selected as the local reference point. The local height axis is defined by the local normal direction passing through that point. In relation to Establish a local rectangular coordinate system within the perpendicular tangent plane. ,in and The two mutually perpendicular transverse coordinate axes defined within the tangential plane are denoted as the kinematic solid-state coordinate system, used to describe the geometric morphology of the particles within this local region; in the... In a plane, using a local reference point Centered on a circle with a preset surface radius Construct a circular patch entity domain for the radius ,in, The value is determined based on the particle's equivalent radius. The instrument's field of view size is determined by setting an upper limit for the patch size based on the average particle diameter or equivalent radius. This ensures that the circular patch only covers a local area of ​​the particle surface without crossing the overall particle scale. Based on the instrument's nominal effective field of view size and the actual imaging coverage, the maximum radius of the circular area that can be accommodated without field-of-view edge clipping and while ensuring data integrity is determined. The smaller of the upper limit radius given by the particle diameter and the upper limit radius given by the instrument's field of view is taken as the maximum radius. The value or upper limit of the value is determined to ensure that the circular patch is spatially small enough to represent a local area, yet completely contained within the effective field of view; the height field data is then used to determine the value of the height field. By mapping to the kinematic solid coordinate system through coordinate transformation, we obtain... Local height function of the independent variable Extract the entity domain that falls into the circular facet. Height data within the circular patch solid domain. Inside, to Low-order shape fitting is performed, preferably using spherical fitting, quadratic surface fitting, or other polynomial surface fitting of at least first and no more than third degree, to obtain the trend surface describing the curvature of the macroscopic shape of the particles. ;exist The residual height is calculated by performing a difference operation between the original height and the trend surface height. and the circular facet entity domain and each of them The point set together serves as the entity of the residual height field, with reference to Figure 3 Through the above processing, the overall macroscopic spherical or ellipsoidal curvature of the particles is separated from the measured morphology, leaving only the height component that reflects the microscopic roughness undulations, providing a physically clear height field basis for subsequent multi-scale roughness analysis.

[0075] Step S2021 constructs a moving solid-state coordinate system and performs curvature stripping. The physical basis for this is that fly ash particles from coal-fired boilers exhibit spherical or ellipsoidal shapes on a macroscopic scale. This invention focuses on the hindering or enhancing effect of the micron-level micro-roughness texture superimposed on their surface on liquid phase wetting. If the original measurement data is used directly, the macroscopic curvature (low-frequency geometric information) of the particles themselves is numerically much larger than the micro-undulations (high-frequency texture information), completely masking the true roughness characteristics. Therefore, this step establishes a moving coordinate system based on local normal vectors and fits and subtracts the macroscopic trend surface within a circular patch domain, mathematically equivalent to virtually flattening the curved particle surface. This differential processing not only eliminates the interference of random posture and macroscopic shape of particles in the measurement space but also extracts the pure residual height field that truly determines the liquid bridge pinning effect, providing a pure geometric reference for subsequent high-precision fractal calculations.

[0076] like Figure 4 The diagram shown is a schematic of multi-scale ring-band Boolean operation and filtering logic.

[0077] Reference Figure 4 The outermost dashed rectangle in the diagram represents the projected measurement area. This refers to the effective field of view boundary of the instrument; the dashed circle inside the diagram represents the solid domain of a circular surface. The black dot in the center of the diagram is the reference point. The gray circular area in the lower right corner of the diagram represents the invalid quality area. The areas outside the center, such as bad spots / shades, are represented by green rings and red dashed arcs that indicate different scales of analysis.

[0078] For specific multi-scale annular zone construction and geometric Boolean operations, refer to Figure 4 Using the moving solid coordinate system The analysis of the residual height field within the solid is described using the example of solid analysis. For the scale k in the middle of the diagram (corresponding to the annular zone)... (i.e., the light gray ring area in the diagram), with reference point Construct concentric ring-shaped core regions with the circle as the center. For example... Figure 4As shown, a gray-shaded area indicates an invalid data region in the lower right of the ring's coverage area, such as invalid data caused by reflective saturation or obstruction. The system performs a geometric Boolean intersection operation. Physically, invalid regions in the gray shaded area of ​​the image were forcibly removed, retaining only the valid sectors in the green area as valid data points at this scale; at this point, the percentage of their valid arc length was calculated. Because it satisfies If the value is greater than or equal to λ, where λ is a preset threshold for the proportion of arc length, it is considered valid, and these green sectors are included in the set of valid sectors.

[0079] For the outermost scale m (red dashed line) in the diagram, when performing arc length percentage threshold filtering, the radius of this ring is relatively large, causing its upper and lower parts to extend beyond the projected measurement area. The boundary is truncated. At this point, the system performs a completeness filter based on the arc length percentage threshold λ: calculating the total arc length percentage of the remaining sectors at this scale. Due to the truncation of the field of view boundary, the amount of effective data is insufficient, resulting in… If the value is less than λ, the system determines that this scale is too affected by boundary effects and the data is incomplete. Therefore, this entire scale is removed and not included in subsequent physical scaling calculations. This represents the effective arc length percentage at scale m. The above process involves the collaborative work of a three-dimensional topography measurement device and a data processing module. Spatial geometric analysis ensures that subsequent fractal calculations are based solely on complete and realistic physical samples.

[0080] Step S2022: Construct a multi-scale annular core domain using the residual height field entity, perform geometric Boolean operations with the projected measurement area and the mass mask to obtain an effective annular domain, and filter the effective sector set according to the arc length ratio threshold.

[0081] In the motion solid coordinate system and circular facet solid domain Inside, refer to Figure 4 , with reference point Define a monotonically increasing sequence of radii centered at a circle. , The scale index is a non-negative integer. This radius sequence can be generated using an arithmetic or geometric progression. For example, it can be set as follows: To be the minimum analysis scale, it is preferably determined based on the measurement spatial resolution or pixel size, such that within a radius of... The smallest neighborhood contains at least a number of valid measurement points; The radius step size between two adjacent scales can be set to a constant step size (arithmetic sequence) according to the preset multi-scale analysis requirements, ensuring that the maximum radius does not exceed the preset patch radius. Thus, the sequence is obtained. Using two adjacent radii and As inner and outer boundaries, in Concentric ring-shaped core domains constructed in a plane This core domain physically corresponds to a distance of [distance] from the reference point on the particle surface. A ring-shaped region within the range forms a multi-scale annular core domain family. The collected projection measurement area is in Defined in a plane This area is defined by the field of view boundary provided by the instrument, the start and end positions of the scan, or the lens cover. It can be determined by taking the intersection of the nominal field of view, the actual imaging edge, and the effective pixel range. This ensures that the areas involved in the analysis are all within the stable imaging range of the equipment. Based on the quality information of the measurement data, such as signal-to-noise ratio, reflection saturation, shadow occlusion, and interference fringe reliability, a quality mask domain is constructed. :exist Internally, a single or a few quality indicators are calculated for each measurement point, such as using only the high signal-to-noise ratio or grayscale contrast as the primary quality criterion. A threshold is set, and measurement points with quality indicators below the threshold are marked as invalid points, while the rest are considered valid points. The results are then calculated from all valid points. A set in a plane is defined as a mass mask domain. This indicates a reliable data range that can be used for subsequent statistical analysis. (Refer to...) Figure 4 .

[0082] For each scale , will ring core domain With circular patch entity domain Projected survey area and quality mask domain Perform geometric Boolean intersection operations to obtain the effective annular zone domain at this scale. Only Internal residual height points are considered valid data points for fractal analysis. Considering that occlusion, edge clipping, or substandard quality can create gaps in the annulus, [the following is considered]. Connectivity analysis along the polar angle direction decomposes the data into several consecutive effective sectors. ,Right now ,in For the first The number of effective sectors at each scale, where ⋃ represents the union operation of sets; for each scale Calculate all valid sectors The sum of the total arc lengths on the ring, and the length of the ideal complete circle. By comparison, the arc length ratio is obtained. ,in This represents the average radius of the annulus corresponding to this scale in the radial direction, used to represent the characteristic length of this scale, and can be taken as... Or other representative radii, Indicates valid sector The geometric arc length, The azimuth sector is numbered to distinguish different azimuth sectors at the same scale; a pre-set arc length percentage threshold is used. For example, 0.6 to 0.8, when At that time, the scale All corresponding valid sectors When included in the set of valid sectors In such cases, the scale is considered to be excessively affected by missing boundaries or incomplete data and is therefore excluded, referring to... Figure 4 Through the above-mentioned multi-scale ring construction, geometric Boolean operation and arc length ratio screening, an effective set of sectors is formed that meets both the instrument field of view and data quality requirements, providing spatial support for subsequent physical scaling calculations.

[0083] Step S2022 constructs multi-scale rings and performs Boolean operations and arc length filtering. The statistical basis for this is that actual industrial microscopic 3D scanning is inevitably affected by instrument field-of-view boundaries, particle occlusion, and high-brightness reflection artifacts, resulting in spatially discontinuous and defective measurement data. Simply performing global statistics would introduce zero or outlier values ​​from invalid regions, leading to severe biases in the structure function calculation. This step uses geometric Boolean operations to forcibly eliminate all unreliable regions and introduces an arc length percentage threshold as a "statistical confidence" criterion. This design is based on the law of large numbers, meaning that only when the amount of effective data in a certain scale ring reaches a certain proportion does the statistical characteristic at that scale have physical representativeness. This rigorous spatial filtering ensures that each subsequently calculated roughness index is based on complete and real physical samples, rather than computational artifacts caused by missing data.

[0084] Step S2023: Perform homogeneous sampling constraints within the set of effective sectors, calculate the height difference structure function that satisfies the point-to-point spacing constraints, and generate physical scales corresponding to each scale;

[0085] For each scale Only in the category of Each effective sector Internal selection of sample point pairs ,in and All are discrete sampling points in the residual height field entity, carrying Coordinates; to ensure data consistency, the original height data is pre-marked according to the scanning method of the 3D measuring device, based on the scan strip number, camera frame number, or laser scanning plane number. A homogeneous sampling constraint is applied when selecting sample point pairs, meaning that sampling only occurs when points... and points Only samples originating from the same scanning trajectory, the same image frame, or the same laser scanning plane are allowed to form valid sample point pairs. This is to avoid interference from roughness calculations caused by splicing errors of different scanning strips or time drift. (Regarding scale...) The spatial spacing between the points must satisfy ,in This indicates that there is a relationship between sample point p and sample point q. Euclidean distance in the plane The point-to-point spacing tolerance is used to allow for a finite discrete error. This applies to all point pairs that satisfy the above-mentioned origin and spacing constraints. Calculate physical scalars: Calculate the height difference structure function In this formula, and These represent the sampling points. and The residual height value corresponding to the residual height field entity; symbol This indicates the ensemble average, with the subscript indicating that the average is applied to all samples whose spatial interval is approximately equal to the analytical radius. The calculation is performed on the effective point pairs. Through this calculation, the analysis radius is obtained. A measure of the intensity of surface undulations. This is achieved by indexing all valid scales. Calculate separately This forms a sequence of physical scalars that varies with scale. This sequence comes directly from the measurable height field and has clear geometric and mechanical physical meanings, laying the foundation for subsequent fractal scaling fitting.

[0086] Step S2023 applies a homogeneous sampling constraint when calculating the height difference structure function. The instrumental basis for this constraint lies in the fact that existing high-precision microscopic three-dimensional topography measurement devices (such as white light interferometry or laser confocal microscopy) typically employ line-by-line scanning, frame-by-frame stitching, or tomographic scanning imaging methods. This means that there are often micro- and nano-level systematic stitching errors or time drifts between different scanning strips or image frames. If sample points are allowed to be selected across strips, these instrumental systematic errors would be misinterpreted as height fluctuations on the particle surface, thus artificially amplifying the roughness value. This step restricts the point pairs used to calculate the height difference to originate from the same scanning trajectory or the same image frame, cutting off the propagation path of instrumental systematic errors at the source. This ensures that the calculated height difference structure function only reflects the inherent physical topography correlation of the particle surface, resulting in measurement results with extremely high robustness and anti-interference capability even at nanometer-level precision.

[0087] Step S2024: Based on the instrument's physical limits and boundary effects, set a clipping threshold, extract an effective scale segment from the physical scale, and perform linear regression fitting in the logarithmic domain to output the particle surface equivalent roughness scaling index.

[0088] Based on the lateral spatial resolution of the three-dimensional measuring device Set minimum analysis scale , It can be obtained by converting the nominal pixel size of the instrument into the magnification of the objective lens, or by measuring known spacing characteristics on a calibration target and converting their pixel spacing in the image into actual length, and is used to characterize the pixel spacing in the image. The minimum distance between adjacent resolvable measurement points on a plane is generally taken as ,in These are empirical coefficients to avoid noise and interpolation errors dominating at excessively small scales; based on the solid domain of circular patches. radius Set maximum analysis radius For example, take ,in It is an empirical coefficient. This is used to avoid significant boundary effects caused by point pairs spanning the entire surface, and to avoid the impact of truncation effects and field-of-view boundary effects near the edges on the statistical results. It is combined with the effective radius set obtained by filtering by arc length proportion. Only the radius is retained. Falling in the range And the effective scale index k corresponding to the height statistics, at each radius Calculate the height difference structure function ,by x-axis Construct data points for the ordinate Perform a linear regression fit on all effective scale points to obtain the slope of the fitted line. In fractal structure function theory, the height difference structure function of a one-dimensional surface profile is usually written as... ,in Indicates the section line at the radius The height difference statistics below, The Hurst index measures the strength of the correlation between height and scale. Taking the logarithm, we get Therefore, in log-log coordinates, the slope satisfies Thus obtain In commonly used surface fractal models, the fractal dimension of a rough surface can be approximated as: ,in A constant related to spatial dimension; Substituting, we can obtain Based on this, this embodiment uses the equivalent roughness scaling index of the particle surface. Treated as a scaling index of the same type as fractal dimension, the presupposed linear conversion relationship is: ,in A pre-defined constant used to measure the slope The range of values ​​is converted into the numerical range of the roughness scaling index; for particle surfaces in three-dimensional space, the range is taken with reference to the three-dimensional surface fractal model. Resulting in explicit calculation formula Through the above process, the height difference structure function is first used within the effective scale range. The slope was obtained by double logarithmic fitting. Then substitute into the linear conversion relationship The equivalent roughness scaling index of the particle surface can then be obtained. This enables the engineering-based quantitative characterization of multi-scale roughness features on particle surfaces.

[0089] Step S2024 sets the clipping threshold and performs log-linear regression. Its mathematical and physical basis lies in the self-similarity theory of fractal geometry and measurement limit constraints: the fractal characteristics of a physical surface do not hold true on infinite scales, but only follow a power-law distribution within a specific "scale-free interval." This step sets a minimum analysis scale to avoid high-frequency noise limited by instrument resolution, and a maximum analysis radius to avoid the finite-size effect region at the edge of the surface patch, thus accurately locking onto the true "linear scaling segment." The slope extracted within this interval directly corresponds to the Hurst exponent, which is then converted into an equivalent roughness scaling exponent through analytical relationships. This process collapses the originally chaotic massive morphological data into a single, physically meaningful quantitative index. This index directly and numerically reveals the spatial filling capability of the particle surface microstructure, providing a solid mathematical interface for accurately calculating the true wetting area of ​​the liquid phase on rough surfaces.

[0090] Step S203: Based on the thickness of liquid phase precipitation on the particle surface, construct the liquid bridge volume calculation equation through the particle surface equivalent roughness scaling index and the liquid phase surface tension coefficient, solve the effective wetting area and liquid bridge volume of the liquid phase in the contact area of ​​the ash particles, and generate a set of liquid bridge physical viscosity characteristic parameters.

[0091] Based on the current local flue gas temperature and the chemical composition of the molten ash phase, determine the surface tension coefficient of the liquid phase at this temperature. and dynamic viscosity The specific methods for determining the surface tension coefficient and dynamic viscosity of silicate melts under different temperatures and compositions are existing technologies in this field and will not be elaborated here. For example, they can be obtained by consulting high-temperature melt property databases or by using general empirical formulas for silicate systems. A calculation model for the effective wetting area considering the influence of micro-roughness is constructed as follows: First, based on the assumption of smooth spherical contact, given the equivalent radius of the gray particles... and the average thickness of the liquid film on the particle surface Under the condition that the liquid film thickness is much smaller than the particle radius, the macroscopic geometric contact radius is calculated. can Approximately expressed as ,in This is an empirical coefficient, which can be taken as [value missing] in typical sphere-to-sphere contact geometry. Approximately equal to The macroscopic wetting radius of the liquid film formed in the contact area when two particles come into contact is used to characterize the contact area; based on this, the nominal contact area is obtained. This area reflects the geometric contact area of ​​the liquid bridge projected onto an ideal smooth sphere between two particles, neglecting microscopic roughness. It utilizes the particle surface equivalent roughness scaling index. To make a microscopic correction to the nominal contact area, a fractal roughness amplification effect is introduced. The correction formula is as follows:

[0092]

[0093] in, This refers to the minimum analysis scale set in step S2024, i.e., the resolution limit determined by the spatial resolution of the three-dimensional topography measurement instrument. In this invention, the physical meaning of this formula is: when... At that time, the micro-fractal roughness structure of the particle surface significantly increased the actual wettable surface area of ​​the liquid phase, making... Greater than the nominal contact area The above index The physical basis comes from the area-scale relationship in fractal geometry. The topological dimension of an ideal smooth sphere is 2, while the equivalent roughness scaling index of a granular surface with micro-texture is... Between 2 and 3, This dimensional difference quantifies the difference in space-filling ability between a rough surface and a two-dimensional smooth projection surface, indicating that the more complex the microscopic rough structure of the particle surface, the better. The larger the area, the more significant the increase in the effective real area that can be wetted by the liquid phase at the microscopic scale relative to the nominal contact area. Based on the principle of volume conservation, assuming that the liquid film near the contact area of ​​gray particles preferentially aggregates into the interparticle gaps and forms a stable liquid bridge under the drive of capillary pressure difference, the initial liquid phase film volume on the particle surface is approximated as all converging in the effective wetting region, and the liquid bridge volume calculation equation is constructed as follows:

[0094]

[0095] That is, through the effective wetting area The volume of the liquid bridge can be calculated by multiplying it by the average liquid film thickness. This volume approximation assumes that the neck region of the liquid bridge formed between the two particles is on the same scale as the average film thickness on the particle surface, and is used in engineering to characterize the total volume of liquid phase contained in the liquid bridge. The calculated geometric parameters (effective wetting area) are then used... Liquid bridge volume ) and the aforementioned physical property parameters (surface tension coefficient) Dynamic viscosity The liquid bridge is co-encapsulated and defined as a set of physical viscous characteristic parameters. This set includes at least area and volume parameters to characterize the capillary suction scale of the liquid bridge, and physical property parameters to characterize the interfacial tension and flow damping capacity of the liquid bridge. The above-mentioned set of physical viscous characteristic parameters of the liquid bridge completely provides the geometric and physical boundary conditions for the morphology and rheological behavior of the liquid bridge at the instant of particle contact. It is passed as the output of this step to step S204 for subsequent calculation of particle adhesion strength based on the improved Rumpf theory.

[0096] Step S204: Based on the set of physical viscosity characteristics parameters of the liquid bridge, the improved Rumpf particle adhesion theory model is used to calculate the resultant force of capillary force, viscous dissipation force and van der Waals force of the liquid bridge at the moment of particle contact, and generate the equivalent adhesion strength distribution on the particle surface.

[0097] An improved Rumpf particle adhesion theory model decouples the transient adhesion between moistened ash particles into three physical components, and performs calculations using a set of liquid bridge physical viscosity parameters and particle motion state data. Specifically, it calculates capillary force. : Utilizing the volume of liquid bridges in the set As a key geometric constraint, the filling angle of the liquid bridge on the particle surface is calculated. Specifically, the geometric volume formula based on the liquid bridge between spheres, such as the approximate relationship... ,in The geometric constant is given by the volume of the liquid bridge. Reverse calculation of the fill angle ;Will Substituting into the capillary force formula with filling angle correction To solve, where The contact angle of the liquid phase on the surface of ash particles is obtained from a wettability database based on the chemical composition (acid-base ratio) of the combustion ash and the type of liquid phase, or extrapolated from the contact angle experimental curve of high-temperature molten slag. The calculation is performed because capillary force essentially acts on the gas-liquid-solid three-phase contact line, and This directly determines the length of the contact line, i.e., the radius of the liquid bridge neck. This method ensures that the calculated adhesion force accurately reflects the physical constraints under the current liquid phase precipitation level, avoiding the overestimation error caused by traditional models that ignore liquid volume limitations. (Calculation of viscous dissipation force) Based on the Stefan-Reynolds lubrication film theory, the calculation formula is as follows: ,in Let be the dynamic viscosity of the liquid phase. The velocity component representing the relative approach of the center-of-mass velocity vectors of the two particles along the line connecting their centers at the moment of collision is denoted as . This velocity value is obtained by real-time analysis of the instantaneous motion vectors of the particles. h is defined as the instantaneous dynamic liquid film thickness between the surfaces of the two particles, a variable that varies with time. To avoid the occurrence of h in numerical calculations... The mathematical singularity of time, setting a physical cutoff thickness , For example, 1–10 nm or according to A certain percentage is set, and when the calculated instantaneous dynamic liquid film thickness is less than this value, it is forcibly taken in the formula. This characterizes the limiting fluid damping state at rough peak contact. The above formula reflects that, under the same viscosity and contact area, the higher the particle separation velocity or the thinner the liquid film, the stronger the anti-separation viscous resistance generated by fluid damping, thereby increasing the equivalent adhesion between particles. Calculation of van der Waals forces. The Hamaker micro-force formula, applicable to liquid / ash particle environments, is adopted. ,in The effective Hamaker constant is calculated using Lifshitz theory and is jointly determined by the Hamaker constant of the ash particle solid phase and the Hamaker constant of the liquid medium. The effective Hamaker constant is constructed based on Lifshitz theory. The specific calculation method is a well-known technology in the field, so the derivation process will not be elaborated here. It is used to characterize the strength of electromagnetic attraction between molecules at extremely small spacing. This force term mainly provides substrate adsorption in extremely thin liquid films or local solid-solid close contact areas, and is particularly important for contact points where liquid bridges are extremely thin or where partial dry spots have been exposed.

[0098] By vector superimposing the above three forces, the total adhesive force of the particle pair at the moment of contact is obtained. Based on this, the equivalent adhesion strength of the contact unit is defined, and Divide by the representative bearing area, for example, the representative bearing area is The equivalent adhesion strength of the local particle surface was obtained. For all or statistically representative particle pairs within the target area of ​​the boiler's convective heating surface, repeat the above calculation process to construct an equivalent adhesion intensity distribution field on the particle surface in a spatial coordinate system. This distribution directly reflects the differences in adhesion ability of different locations and particle combinations under the current working conditions, and serves as the input for the subsequent step S300 to drive the prediction of the generation of gray particle agglomerates.

[0099] Step S203 addresses the micron-level texture rich on the surface of biomass ash particles. Instead of using the traditional ideal smooth sphere model, it employs the "area-scale" relationship from fractal geometry, utilizing the roughness scaling index generated in step S202 to physically correct the nominal contact area. This is based on the fact that roughness significantly increases the actual spreadable surface area of ​​the liquid phase (i.e., the effective wetting area), thereby altering the volume distribution of the liquid bridge and the capillary suction reference. Step S204, based on an improved Rumpf adhesion theory, decouples the total adhesion force between particles into three independent physical components for calculation: using... The liquid bridge volume constraint is used to solve for the filling angle to calculate the capillary force (based on the length of the gas-liquid-solid three-phase contact line). The approach velocity is introduced based on the Stefan-Reynolds lubricating film theory to calculate the viscous dissipative force (based on the squeezing damping effect of the fluid in an extremely narrow gap). The van der Waals force is calculated based on the Hamaker constant (based on the electromagnetic attraction between molecules). This method of vector superposition after separate calculations overcomes the shortcomings of traditional empirical formulas that cannot dynamically respond to the particle motion state (such as collision velocity) and microscopic liquid volume changes, thus accurately quantifying the true "adhesive gripper" strength of particles at the moment of contact.

[0100] The intelligent prediction method for alkali metal slagging tendency in biomass gas coupled coal-fired boilers also includes the following steps:

[0101] Step S300: Based on the equivalent adhesion intensity distribution on the particle surface, statistically analyze the particle turbulent collision physical parameters, determine the effective collision adhesion efficiency based on energy competition, calculate the agglomeration growth rate of gray particles in the flow field, and output the spatial growth configuration of gray particle agglomerates.

[0102] The spatial growth configuration of the gray particle aggregates described in step S300 specifically includes the following steps:

[0103] Step S301: Based on the flue geometry of the boiler's convective heating surface and the average flue gas velocity, estimate the collision parameters of ash particles using empirical formulas from turbulence theory, and statistically analyze the particle collision frequency and relative impact velocity.

[0104] Obtain the flue gas flow cross-sectional area of ​​the target region of the convective heating surface. and wetted perimeter Using the formula Calculate the equivalent characteristic scale of the flue The hydraulic diameter, or , is a dimension that equates a non-circular flue to a regular fluid channel; combined with the measured average flue gas velocity... and gas kinematic viscosity Calculate the flow Reynolds number And using classical turbulence correlations Estimating the average turbulent dissipation rate of the flow field , Let be an empirical constant. Based on Stokes' law of drag, calculate the relaxation time of the particles in the flow field. ,in The average density of the ash particles. This represents the average particle size of the initial ash particles produced after the mixed fuels are burned in the furnace. It is usually obtained through coal quality analysis or fly ash particle size analysis. The time interval represents the flue gas density and characterizes how quickly particles respond to changes in fluid velocity; the turbulence intensity coefficient is determined. This coefficient can be taken as 0.05 to 0.1 based on experience with fully developed pipe flow, or it can be determined by an empirical formula. The root mean square value of the estimated turbulent fluctuation velocity was calculated. ; Calculate the Stokes number of the particles ,in The Kolmogorov microscale time is used to characterize the inertial following effect of particles in turbulent eddies. Based on the above parameters, the statistics are directly calculated using the Abrahamsen inertial collision theory model: the particle number density, and Substituting into the Abrahamson flux formula, the particle collision frequency is calculated. The specific formula is as follows:

[0105]

[0106] in, Particle number density, which is the number of particles per unit volume. This is the collision geometry constant, typically taken as 5.0 or higher based on the isotropic turbulence assumption. , The inertial response correction parameter is, for example, set to 0.5~0.65. This formula, based on Abrahamson's inertial collision theory, explicitly introduces the Stokes number to correct the underestimation of the collision rate in the traditional Saffman-Turner formula under high Reynolds number and high inertial particle conditions. This allows it to more accurately reflect the frequent encounter behavior of fly ash particles in strong turbulence under the conditions described in this invention. Based on statistical dynamics... ( The relative impact velocity of the particle pair at the moment of collision is calculated using a model constant or a similar mean square relative velocity formula. This implementation method is based on the classical Kolmogorov turbulence theory and Abrahamsen inertial collision theory, and has been widely used in engineering for rapid estimation of gas-solid two-phase flow. It can effectively avoid the high cost of large eddy simulation, thereby meeting the real-time requirements of intelligent prediction systems.

[0107] Step S302: Calculate the particle collision kinetic energy based on the relative impact velocity, and perform a competition analysis between the particle collision kinetic energy and the critical adhesion energy converted from the equivalent adhesion intensity distribution on the particle surface to calculate the effective collision adhesion efficiency.

[0108] Based on relative impact velocity Calculate the equivalent mass of the gray particles involved in the collision. The specific method for obtaining the density is as follows: the average density of ash particles is obtained based on the coal quality analysis data entering the furnace. Combined with the average particle size of the original ash particles Using the formula for the volume of a sphere The characteristic collision kinetic energy of the particle swarm under the current turbulence intensity was calculated. This index statistically represents the average level of inertia at which particles attempt to bounce and separate when they collide with each other in a flow field.

[0109] The critical adhesion energy threshold is constructed by using the equivalent adhesion strength distribution on the particle surface. This strength value integrates the capillary force of the liquid bridge, the viscous dissipation force, and the van der Waals force; combined with the effective wetting area. ,Will It is considered as the anti-separation power density on the contact surface and multiplied by the representative separation displacement. In this embodiment, the liquid phase precipitation thickness on the particle surface calculated in step S2012 is directly selected. As a representative separation displacement ,Right now The physical basis for this is that the fracture length of a liquid bridge is usually positively correlated with the thickness of the liquid film; the thicker the liquid film, the longer the stretching distance the liquid bridge can withstand during particle separation. Based on this, the critical adhesion energy required for a single effective collision can be calculated. This threshold represents the maximum rebound energy that can be overcome by the formation of stable agglomeration between particles under the current liquid phase precipitation state.

[0110] Calculating the effective collisional adhesion efficiency: Based on the principle of energy competition, the critical capture velocity for particle capture is deduced from the critical adhesion energy. Since step S301 outputs the root mean square velocity in a statistical sense, this step uses the probability integral method to calculate the efficiency: it is assumed that the relative collision velocities between particles in the turbulent field follow a Maxwell-Boltzmann distribution or other probability density functions that conform to statistical thermodynamics. The function is The characteristic velocity scale parameters are constructed, for example, using the Maxwell-Boltzmann distribution form. ,in To and The directly related velocity dispersion parameter is given by the following formula: Let s represent the instantaneous relative velocity of any pair of particles at the moment of collision; the specific execution of the probability integral method is as follows: for this probability density function in the interval... Perform definite integral operations on the above, that is, calculate The integral result represents the proportion of particles in the particle swarm whose relative velocity is less than the critical capture velocity; this proportion is denoted as the effective collision adhesion efficiency. This efficiency quantifies the proportion of physical collisions occurring in turbulence under the current operating conditions that can be transformed into irreversible aggregation and adhesion, providing accurate source term inputs for subsequent population equilibrium models.

[0111] Step S303: Based on the particle collision frequency and effective collision adhesion efficiency, the degree of agglomeration growth of gray particles in the flow field is calculated by the characteristic agglomeration rate integral method, and the spatial growth configuration of gray particle agglomerates is generated.

[0112] Multiplying the particle collision frequency by the effective collision adhesion efficiency yields the effective agglomeration frequency. This multiplication is based on the principle of independent events in statistics, meaning that an effective agglomeration event must simultaneously satisfy two independent conditions: "collision occurs" and "energy competition is won." Therefore, the product of these two conditions physically accurately represents the average rate at which a single ash particle successfully captures other particles and forms a cluster per unit time. The agglomeration multiplication factor is then calculated to obtain the average flue gas residence time of ash particles in the target area of ​​the convective heating surface. The time span through the characteristic length of the flue in the target area Divide by the average flue gas velocity Based on the principle of kinetic integration, the effective aggregation frequency is obtained. With average flue gas residence time Multiply by this to get the theoretical number of collisions. This parameter quantifies the total number of effective adhesion events that occur cumulatively throughout the particle's entire lifecycle as it traverses the region; it also calculates the average particle size of the aggregates. An exponential volume growth model is adopted, expressed by the formula as follows: The basis for this calculation logic is that, in the "snowball" effect, the volume of agglomerates increases cumulatively with the number of captures, and the particle size is a cube root function of the volume. This formula can quickly estimate the final equivalent size of fly ash before it falls into the cold ash hopper with low computing power; determining the fractal dimension and porosity, based on the effective collision adhesion efficiency... Determine the fractal dimension The physical basis for this judgment comes from the "restricted aggregation mechanism" in colloid science. An empirical threshold is set, for example, 0.5. When the particle size is such that it sticks together upon contact, it belongs to "ballistic aggregation," forming an extremely loose and dendritic structure. In this case, a lower fractal dimension is used, such as... ;when When the particle size is high, it indicates that the particles need to undergo multiple collisions and adjustments to adhere, which is a type of "reaction-limited aggregation." The structure is relatively dense, and a higher fractal dimension is used in this case. Substitute into the fractal porosity formula The output includes the average particle size of the aggregates. With high porosity The spatial growth configuration of ash particle aggregates was obtained, thus numerically reproducing the characteristics of biomass ash flocs that are large in volume but loose in structure.

[0113] Steps S301 to S303 do not simply assume that collision equals agglomeration. Instead, they first use Abrahamson's inertial collision theory and Kolmogorov's microscale turbulence theory to statistically determine the collision frequency and relative impact velocity of particles in the flow field, thus constructing the rebound kinetic energy of the particles. An energy competition criterion is introduced to analyze the adversarial relationship between the critical adhesion energy of the particle surface (determined by the liquid bridge strength and liquid film thickness, representing the trapping ability) and the particle collision kinetic energy (representing the escape ability). The effective collision adhesion efficiency is calculated using probability density integration, which physically explains why only a portion of collisions can form effective agglomeration. When calculating the agglomerate growth configuration, the concept of restricted agglomeration from colloid science is introduced. The fractal dimension of the agglomerate is dynamically determined based on the adhesion efficiency: high adhesion efficiency leads to a dendritic, loose structure that adheres easily upon contact (low fractal dimension), while low efficiency leads to a dense, spherical structure (high fractal dimension). This process fully reproduces the dynamic evolution trajectory of ash particles from monomers to loose flocs, providing an accurate geometric precursor for subsequent prediction of the formation of large, light slag masses in the ash hopper.

[0114] The intelligent prediction method for alkali metal slagging tendency in biomass gas coupled coal-fired boilers also includes the following steps:

[0115] Step S400: Based on the spatial growth configuration of the ash particle agglomerates, perform macroscopic mechanical characteristic analysis and bridging stability determination on the accumulation body at the throat of the cold ash hopper, and output the ash hopper blockage risk level;

[0116] Step S400 involves macroscopic mechanical characteristic analysis and bridging stability determination of the accumulation at the throat of the cold ash hopper, including the following steps:

[0117] Step S401: Based on the spatial growth configuration of the aggregates and combined with the thickness of liquid phase precipitation on the particle surface, the physical properties of the accumulation body at the throat of the cold ash hopper are estimated and analyzed in an engineering manner to obtain the macroscopic equivalent density and overall shear strength.

[0118] Determine the bed porosity , here It is based on the fractal dimension of aggregates An inverse proportional mapping relationship is constructed using a linear interpolation function: a lower limit for the bed porosity corresponding to the dense sphere is set. (e.g., 0.4) and the upper limit of bed porosity corresponding to loose dendrites. (e.g., 0.6), based on the current In typical range Internal position calculation This relationship reflects that the more irregular the shape of the aggregate ( The lower the porosity, the looser the packing; this is achieved using the dual porosity reduction formula. The macroscopic equivalent density was calculated. Derivation of the overall shear strength of the accumulation body The physical derivation logic of this strength is as follows: The equivalent adhesion strength distribution on the particle surface... As a baseline, multiply by a coefficient characterizing the effective solid contact density within the packing. Furthermore, a sintering enhancement correction based on the time dimension is introduced; specifically, the slag discharge cycle of the slag discharger is used as the static residence time in the calculation. Combined with liquid phase precipitation thickness Constructing sintering strengthening factors ,in The sintering rate constant is related to the liquid phase viscosity. The time hardening exponent is calculated as follows: This step accurately quantifies the macroscopic structural strength formed by the combined effects of liquid bridge contact density and time-accumulated sintering in the ash hopper on loose ash.

[0119] Step S402: Based on the macroscopic equivalent density and overall shear strength, construct the competition analysis logic for bridge stability and determine the ash hopper blockage risk level;

[0120] Constructing a dimensionless bridging tendency index This index, based on a simplified mechanical equilibrium model of Jenike's granular material flow theory, aims to compare the "bond strength that maintains the stability of the arch bridge" with the "gravity load that destroys the arch bridge." Its calculation formula is as follows: ,in For overall shear strength, For macroscopic equivalent density, It is the acceleration due to gravity. The actual geometric width of the throat of the cold ash hopper is determined before calculation. , as well as Maximum-minimum normalization is performed to map it to a dimensionless value of 0-1; this index intuitively reflects the high-risk bridging characteristics of biomass ash due to its extremely low density (small denominator) and hardening during sintering (large numerator); risk assessment is performed, and the calculated... With bridging critical threshold Comparison: the bridging critical threshold The risk boundary value was determined by performing binary statistical regression analysis on calculated indices under normal ash discharge and bridging blockage conditions in the boiler's historical operating database; if If the current operating condition is deemed high-risk, and a large blockage of light slag is predicted, the system will output a severe slag formation warning signal; if If the risk is low, the slag discharge is predicted to be smooth, thus achieving intelligent prediction of slag formation tendency.

[0121] To address the unique characteristics of biomass ash—its large volume, light weight, and susceptibility to low-temperature sintering—step S401 establishes a dual porosity model. This model subtracts the microscopic pores within the agglomerates and the macroscopic voids in the bed, theoretically deriving an extremely low macroscopic equivalent density. Simultaneously, time-hardening index and liquid phase viscosity parameters are introduced to quantify the strength evolution caused by liquid bridge fusion during settling. Step S402, based on the Jenike mechanical equilibrium model, constructs a dimensionless bridging tendency index. Its core physical significance lies in comparing the bond strength maintaining the arch bridge with the gravitational load destroying it. By normalizing key parameters to eliminate dimensional differences, this index keenly captures the extreme bridging risk arising from the contradictory characteristics of insufficient collapse driving force due to the low density of biomass ash and excessive structural strength caused by sintering. This provides accurate early warning of large-scale light slag blockage in the cold ash hopper.

[0122] Example 2

[0123] like Figure 2 As shown, the intelligent prediction system for alkali metal slagging tendency in biomass gas coupled coal-fired boilers includes the following:

[0124] The data acquisition module is used to acquire flue gas temperature data of the target area of ​​the boiler's convective heating surface, and to acquire ash melting characteristic temperature and alkali metal content data of the mixed fuel entering the furnace.

[0125] The adhesion distribution module is used to perform physically constrained adaptive correlation calculation of ash temperature-liquid phase yield based on the flue gas temperature data, ash melting characteristic temperature and alkali metal element content data, construct a set of liquid bridge physical viscosity characteristic parameters in combination with particle surface micro-roughness characteristics, calculate the coupling adhesion resultant force, and generate the equivalent adhesion strength distribution of particle surface.

[0126] The aggregate space module is used to statistically analyze the particle turbulent collision physical parameters based on the equivalent adhesion intensity distribution on the particle surface, determine the effective collision adhesion efficiency based on energy competition, calculate the agglomeration growth rate of gray particles in the flow field, and output the spatial growth configuration of gray particle agglomerations.

[0127] The risk assessment module is used to perform macroscopic mechanical characteristic analysis and bridging stability assessment on the accumulation body at the throat of the cold ash hopper based on the spatial growth configuration of the ash particle agglomerates, and output the ash hopper blockage risk level.

[0128] It should be understood that although the steps in the flowcharts of the various embodiments of the present invention are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the various embodiments may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least a portion of the sub-steps or stages of other steps.

[0129] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0130] The foregoing description is illustrative of the invention and should not be construed as limiting it. Although several exemplary embodiments of the invention have been described, those skilled in the art will readily understand that many modifications can be made to the exemplary embodiments without departing from the novel teachings and advantages of the invention. Therefore, all such modifications are intended to be included within the defined scope of the invention. It should be understood that the foregoing description is illustrative of the invention and should not be construed as limiting it to the specific embodiments disclosed, and modifications to the disclosed embodiments and other embodiments are intended to be included within the appended scope of protection. The invention is defined by its scope of protection and its equivalents.

Claims

1. A smart prediction method for alkali metal slagging tendency in biomass gas coupled coal-fired boilers, characterized in that, include: Acquire flue gas temperature data for the target area of ​​the boiler's convective heating surface, and acquire ash melting characteristic temperature and alkali metal content data for the mixed fuels fed into the furnace. Based on the flue gas temperature data, ash melting characteristic temperature and alkali metal element content data, a physically constrained adaptive correlation calculation of ash temperature-liquid phase yield is performed. A set of liquid bridge physical viscosity characteristic parameters is constructed by combining the micro-roughness characteristics of particle surface, the coupling adhesion resultant force is calculated, and the equivalent adhesion strength distribution of particle surface is generated. Based on the equivalent adhesion intensity distribution on the particle surface, the physical parameters of particle turbulent collision are statistically analyzed, the effective collision adhesion efficiency based on energy competition is determined, the agglomeration growth rate of gray particles in the flow field is calculated, and the spatial growth configuration of gray particle agglomerates is output. Based on the spatial growth configuration of the ash particle agglomerates, macroscopic mechanical characteristics analysis and bridging stability determination are performed on the accumulation in the throat of the cold ash hopper, and the ash hopper blockage risk level is output. The step of generating the equivalent adhesion strength distribution on the particle surface is as follows: Based on the flue gas temperature data, ash melting characteristic temperature and alkali metal element content data, a physically constrained adaptive correlation calculation of ash temperature-liquid phase yield is performed to determine the thickness of liquid phase precipitation on the particle surface. Based on the three-dimensional topography measurement data of particle surfaces, a residual height field is constructed by curvature stripping. The height difference structure function is calculated in the multi-scale effective analysis domain. Log-linear regression analysis is used to generate the equivalent roughness scaling index of particle surfaces. Based on the thickness of liquid phase precipitation on the particle surface, an equation for calculating the liquid bridge volume is constructed using the equivalent roughness scaling index of the particle surface and the surface tension coefficient of the liquid phase. Based on the liquid bridge volume calculation equation, the effective wetting area and liquid bridge volume of the liquid phase in the contact area of ​​ash particles are calculated, and a set of liquid bridge physical viscosity characteristic parameters is generated. Based on the set of physical viscosity characteristics parameters of the liquid bridge, an improved Rumpf particle adhesion theory model is used to calculate the resultant force of capillary force, viscous dissipation force and van der Waals force of the liquid bridge at the moment of particle contact, and generate the equivalent adhesion intensity distribution on the particle surface. The step of generating the spatial growth configuration of the gray particle agglomerates includes: Based on the flue geometry of the boiler's convective heating surface and the average flue gas velocity, the collision parameters of ash particles are estimated using empirical formulas based on turbulence theory, and the particle collision frequency and relative impact velocity are statistically analyzed. The particle collision kinetic energy is calculated based on the relative impact velocity. The particle collision kinetic energy is then compared with the critical adhesion energy, which is transformed from the equivalent adhesion intensity distribution on the particle surface, to calculate the effective collision adhesion efficiency. Based on particle collision frequency and effective collision adhesion efficiency, the characteristic aggregation rate integral method is used to calculate the degree of aggregation growth of gray particles in the flow field, and generate the spatial growth configuration of gray particle agglomerates. The steps for determining the risk level of clogging in the output ash hopper include: Based on the spatial growth configuration of aggregates and combined with the thickness of liquid phase precipitation on particle surface, the physical properties of the accumulation body in the throat of cold ash hopper are estimated and analyzed in an engineering manner to obtain the macroscopic equivalent density and overall shear strength. Based on macroscopic equivalent density and overall shear strength, a competitive analysis logic for bridge stability is constructed to determine the risk level of ash hopper blockage.

2. The intelligent prediction method for alkali metal slagging tendency in biomass gas coupled coal-fired boilers according to claim 1, characterized in that, The steps for determining the thickness of liquid phase precipitation on the particle surface are as follows: The residence dose in the eutectic sensitive temperature zone is calculated using the flue gas temperature data, and the effective alkali metal index is obtained by performing an activity correction on the alkali metal element content data using the ash melting characteristic temperature. The effective liquid volume fraction was calculated by adaptively correcting the preset reference gray temperature-liquid phase yield relationship using the residence dose and effective alkali metal index in the eutectic sensitive temperature zone as physical constraint factors. Based on the particle geometric characteristic parameters, a conversion relationship between volume fraction and film thickness is constructed, and the effective liquid phase volume fraction is converted into the liquid phase precipitation thickness on the particle surface.

3. The intelligent prediction method for alkali metal slagging tendency in biomass gas coupled coal-fired boilers according to claim 1, characterized in that, The step of generating the equivalent roughness scaling index of the particle surface includes: Based on the height field data measured in three dimensions, a servo-driven solid coordinate system and a circular surface solid domain are established. Within the domain, macroscopic curvature is stripped away through low-order shape fitting and differential operations to generate a residual height field solid. The residual height field entity is used to construct a multi-scale annular core domain. Geometric Boolean operations are performed with the projected measurement area and the mass mask to obtain the effective annular domain. The effective sector set is then obtained by filtering based on the arc length ratio threshold.

4. The intelligent prediction method for alkali metal slagging tendency in biomass gas coupled coal-fired boilers according to claim 3, characterized in that, The step of generating the equivalent roughness scaling index of the particle surface further includes: Within the set of effective sectors, perform homogeneous sampling constraints, calculate the height difference structure function that satisfies the point-to-point spacing constraints, and generate physical scales corresponding to each scale. Based on the instrument's physical limits and boundary effects, a clipping threshold is set, and the effective scale segment of the physical scale is truncated. Then, linear regression fitting is performed in the logarithmic domain to output the equivalent roughness scaling index of the particle surface.

5. A biomass gas coupled coal-fired boiler alkali metal slagging tendency intelligent prediction system, used to implement the biomass gas coupled coal-fired boiler alkali metal slagging tendency intelligent prediction method according to any one of claims 1-4, characterized in that, include: The data acquisition module is used to acquire flue gas temperature data of the target area of ​​the boiler's convective heating surface, and to acquire ash melting characteristic temperature and alkali metal content data of the mixed fuel entering the furnace. The adhesion distribution module is used to perform physically constrained adaptive correlation calculation of ash temperature-liquid phase yield based on the flue gas temperature data, ash melting characteristic temperature and alkali metal element content data, construct a set of liquid bridge physical viscosity characteristic parameters in combination with particle surface micro-roughness characteristics, calculate the coupling adhesion resultant force, and generate the equivalent adhesion strength distribution of particle surface. The aggregate space module is used to statistically analyze the particle turbulent collision physical parameters based on the equivalent adhesion intensity distribution on the particle surface, determine the effective collision adhesion efficiency based on energy competition, calculate the agglomeration growth rate of gray particles in the flow field, and output the spatial growth configuration of gray particle agglomerations. The risk assessment module is used to perform macroscopic mechanical characteristic analysis and bridging stability assessment on the accumulation body at the throat of the cold ash hopper based on the spatial growth configuration of the ash particle agglomerates, and output the ash hopper blockage risk level.

Images