System and method for tail gas emission modeling analysis of municipal solid waste incineration process

Abstract

The present invention provides a system and method for tail gas emission modeling analysis of a municipal solid waste incineration process. The system comprises a reference working condition multi-software coupling full-process numerical simulation model construction module, a multi-operation working condition simulation mechanism data acquisition module, a MIMO-LRDT-based tail gas emission model construction module and a single / double-factor-based tail gas emission analysis module. The reference working condition multi-software coupling full-process numerical simulation model construction module is connected to the multi-operation working condition simulation mechanism data acquisition module. The multi-operation working condition simulation mechanism data acquisition module is connected to the MIMO-LRDT-based tail gas emission model construction module. The MIMO-LRDT-based tail gas emission model construction module is connected to the single / double-factor-based tail gas emission analysis module. The system and method for tail gas emission modeling analysis of a municipal solid waste incineration process provided by the present invention can achieve tail gas emission modeling analysis of the municipal solid waste incineration process.

Description

[Corrected 11.05.2024 according to Rule 26] Modeling and analysis system and method for tail gas emissions from municipal solid waste incineration Technical Field

[0001] The present invention relates to the technical field of municipal solid waste incineration, and in particular to a system and method for modeling and analyzing tail gas emissions during a municipal solid waste incineration process. Background Art

[0002] The generation of municipal solid waste (MSW) is increasing with economic development and accelerated urbanization, reaching 3.4 billion tons globally by 2050. MSW incineration (MSWI) technology converts waste into energy (WTE) through fermentation, combustion, heat exchange, and purification processes. Its advantages in harmlessness, waste reduction, and resource utilization have led to its widespread adoption. my country's MSWI technology has largely been imported from developed countries such as Europe and Japan, which employ automatic combustion control (ACC) systems for stable operation. However, the high moisture content, low calorific value, and complex composition of MSW in my country have hindered the practical application of ACC systems in my country. Currently, in industrial settings, domain experts rely on their experience to manually set "air distribution and material distribution" operational variables to address issues such as abnormal furnace temperatures and excessive exhaust emissions, combining multimodal information from distributed control systems (DCSs), flame video images, and team work records. This manual operation mode is subject to significant randomness, lags, limited staff resources, and variability in expert experience, making it difficult to ensure long-term stable operation of MSWI plants. Therefore, it is necessary to study the mapping relationship between the “air and material distribution” operating variables and the exhaust gas emissions to support the control strategy of the MSWI process.

[0003] The MSWI process, a typical process industry, involves complex physical and chemical reactions involving numerous variables that are mutually coupled, making it difficult to construct an accurate mathematical model. Currently, numerical simulation software such as FLIC, computational fluid dynamics (CFD), and Aspen Plus have become effective tools for analyzing the MSWI process due to their efficiency, cost-effectiveness, and ease of use. However, using a single numerical simulation software is difficult to effectively analyze the complex mapping relationship between "air distribution and material distribution" variables and exhaust emission concentrations. Therefore, it is essential to design a system and method for modeling and analyzing exhaust emissions from municipal solid waste incineration processes.

[0004] Summary of the Invention

[0005] The purpose of the present invention is to provide a system and method for modeling and analyzing tail gas emissions from a municipal solid waste incineration process, which can realize modeling and analysis of tail gas emissions from a municipal solid waste incineration process.

[0006] To achieve the above object, the present invention provides the following solutions:

[0007] A system for modeling and analyzing tail gas emissions from a municipal solid waste incineration process, comprising: a multi-software coupled full-process numerical simulation model construction module under a baseline operating condition, a multi-operating operating condition simulation mechanism data acquisition module, a tail gas emission model construction module based on MIMO-LRDT, and a tail gas emission analysis module based on a single / dual factor, wherein the multi-software coupled full-process numerical simulation model construction module under the baseline operating condition is connected to the multi-operating operating condition simulation mechanism data acquisition module, the multi-operating operating condition simulation mechanism data acquisition module is connected to the tail gas emission model construction module based on MIMO-LRDT, and the tail gas emission model construction module based on MIMO-LRDT is connected to the tail gas emission analysis module based on a single / dual factor;

[0008] The multi-software coupled full-process numerical simulation model construction module under the benchmark working conditions is used to construct a full-process numerical simulation model that can fit the industrial field benchmark working conditions;

[0009] The multi-operation condition simulation mechanism data acquisition module is used for the full-process data simulation model to acquire simulation mechanism data under multiple operation conditions;

[0010] The exhaust emission model construction module based on MIMO-LRDT is used to construct an exhaust emission model based on the MIMO-LRDT algorithm;

[0011] The tail gas emission analysis module based on single / double factors is used to perform single / double factor analysis on the mapping relationship between the operating variables and the tail gas emission concentration based on the tail gas emission model.

[0012] The present invention also provides a method for modeling and analyzing tail gas emissions from a municipal solid waste incineration process, which is applied to the above-mentioned system for modeling and analyzing tail gas emissions from a municipal solid waste incineration process, and includes the following steps:

[0013] Step 1: Based on the multi-software coupled full-process numerical simulation model construction module under the benchmark working conditions, a full-process numerical simulation model that can fit the industrial site benchmark working conditions is constructed;

[0014] Step 2: Based on the multi-operating condition simulation mechanism data acquisition module, the simulation mechanism data under multiple operating conditions is acquired through the full-process data simulation model;

[0015] Step 3: Construct an exhaust emission model based on the MIMO-LRDT algorithm through the exhaust emission model construction module based on the MIMO-LRDT algorithm;

[0016] Step 4: Perform single / double factor analysis on the mapping relationship between the operating variables and the exhaust emission concentration based on the exhaust emission model through the exhaust emission analysis module based on single / double factors.

[0017] Optionally, in step 1, a full-process numerical simulation model capable of fitting the industrial site benchmark working conditions is constructed based on a multi-software coupled full-process numerical simulation model construction module under the benchmark working conditions, specifically comprising the following steps:

[0018] Step 101: Solid phase combustion process on the grate based on FLIC simulation;

[0019] Step 102: Simulating the gas phase combustion process in the furnace using Fluent;

[0020] Step 103: Waste heat exchange and other flue gas purification processes based on Aspen Plus simulation.

[0021] Optionally, in step 101, the solid phase combustion process on the grate based on FLIC simulation is specifically as follows:

[0022] The solid-phase combustion process on the grate is simulated using FLIC, which includes the solid-phase MSW combustion model, basic conservation equations, and component transport and heat radiation conversion equations.

[0023] The solid-phase MSW combustion process includes water evaporation, volatile analysis, volatile combustion and coke oxidation stages. MSW is pushed from the feeder to the grate. After passing through the high-temperature heat radiation in the furnace and the furnace wall, the water gradually evaporates at a rate of:

[0024] Where R evp is the water evaporation rate, S a is the particle surface area, h s is the convective mass transfer coefficient, C w,s and C w,g are the moisture content in MSW and mixed gas, Q cr To absorb heat during radiation and convection heat transfer, H evp The heat required for water evaporation is absorbed, T s is the MSW temperature. After the water evaporates completely, the MSW reaches the volatile analysis temperature. The release process of the main volatiles is: MSW→Volatile(C m H n ,CO,CO2,H2O)+Char (2)

[0025] The volatile gas mixes with air and burns. The corresponding combustion reaction is: C m H n +(m / 2+n / 4)O2→mCO+n / 2H2O (3)CO+1 / 2O2→CO2 (4)

[0026] Its combustion rate is the minimum of the kinetic rate and the mixing rate. MSW gradually turns into coke as the volatile matter is released. It further reacts with air to produce CO and CO2, which is: C + αO2 → 2(1-α)CO + (2α-1)CO2 (5)

[0027] Wherein, α is the stoichiometric coefficient, and 0.5≤α≤1.

[0028] Optionally, in step 102, the gas phase combustion process in the furnace simulated by Fluent is specifically as follows:

[0029] The FLIC output is regarded as the boundary condition of the Fluent numerical simulation. In Fluent, the combustion component CmHn is replaced by CH4. The second-order upwind scheme and SIMPLE algorithm are used to discretize and solve the gas phase component combustion equation. The thermal radiation DO model is:

[0030] Where I is the radiation intensity, and are the position vector and direction vector respectively, a is the absorption coefficient, σ s is the diffusion coefficient, n is the refractive index, σ is the Boltzmann constant, Φ is the phase function, Ω′ is the fixed angle, and the standard k-ε model for turbulent gas flow is:

[0031] Where ρ is the gas density, k is the turbulent kinetic energy, and u i is the speed, x i and x j The coordinate system μ is the turbulent viscosity, and the eddy dissipation conceptual model for solving the interaction between gas flow and combustion chemical reaction is:

[0032] Where Y i is the mass fraction of substance i, is the diffusion flux of the substance, R i is the net production rate, S i To generate additional rates, the thermal radiation distribution data after the Fluent simulation is used as the input of FLIC for iterative coupling. When the temperature difference between the two outputs meets the set error, the visualized temperature field, velocity field, and concentration field in the furnace are output.

[0033] Optionally, in step 103, the waste heat exchange and flue gas purification processes simulated based on Aspen Plus are specifically as follows:

[0034] The flue gas temperature and flue gas components output by FLIC, the furnace temperature output by Fluent, the MSW ash content obtained based on laboratory tests, and the secondary air temperature, secondary air flow rate, and urea solution dosage collected from actual industrial sites are used as input. The flue gas components include H2O, N2, O2, H2, CO, CO2, CH4, and S. The Gibbs reactor RDibbs is used to simulate the combustion process in the furnace, and the stoichiometric reactor Rstoic1 is used to simulate the denitrification reaction. The reactions involved are:

[0035] In the heat exchange process stage of the waste heat boiler, two stream heat exchanger modules are used to simulate the heat exchange process between the high-temperature flue gas and the superheater and economizer to obtain the cooled furnace outlet flue gas G1. In the flue gas treatment process stage, a stoichiometric reactor is used to simulate the deacidification process of the flue gas G1. The simulation results are as follows:

[0036] The mixer module is used to simulate the adsorption process of heavy metals, dioxins and other pollutants in the flue gas by activated carbon, and the component separator module is used to simulate the bag filter. Finally, the diverter module is used to simulate the process of fly ash entering the ash bin, thereby obtaining the treated flue gas G2. In the flue gas emission stage, the compressor module is used to simulate the flue gas G3 entering the atmosphere.

[0037] Optionally, in step 2, the simulation mechanism data under multiple operating conditions is obtained through the full-process data simulation model based on the multiple operating condition simulation mechanism data acquisition module, specifically:

[0038] One non-operated variable and 11 operated variables were selected for a four-level orthogonal experimental design, in which the non-operated variable was the MSW component, and the operated variables included feed rate, grate speed, primary air temperature 1, primary air temperature 2, primary air temperature 3, primary air flow 1, primary air flow 2, primary air flow 3, primary air flow 4, secondary air temperature and secondary air flow. Based on the orthogonal experimental design, the simulation mechanism data under multiple operating conditions were obtained through the full-process data simulation model.

[0039] Optionally, in step 3, an exhaust emission model based on the MIMO-LRDT algorithm is constructed by using an exhaust emission model construction module based on the MIMO-LRDT algorithm, specifically:

[0040] Based on the simulation mechanism data under the above multiple operating conditions, a multi-input and multi-output linear regression decision tree algorithm is used to establish an exhaust emission model including CO, CO2, O2, SO2 and NOx emissions. The obtained simulation mechanism data is recorded as

[0041] The mean square error of the target value is calculated by traversing the simulation mechanism data:

[0042] Where, represents the loss function value of MSE, (n,i) represents the i-th eigenvalue of the n-th sample in the k-th iteration;

[0043] Based on the calculation results, select the minimum MSE as the first non-leaf node As the splitting variable, it is:

[0044] Repeat the above process and obtain (T / 2-1) intermediate nodes based on the minimum number of samples θ set by experience

[0045] The linear regression method is used to calculate the predicted output of the CART leaf node, which is:

[0046] Where, and are the output, input and weight matrix of the tth leaf node respectively;

[0047] Obtaining the weight vector is reduced to solving a class of overdetermined matrix equations, and using the regularized least squares loss function, which is:

[0048] In the formula, λ represents the regular term coefficient, and λ≥0. The above loss function is The gradient of is expressed as:

[0049] make have to:

[0050] According to the obtained weight vector, the leaf node prediction value is calculated. Considering all leaf nodes, the MIMO-LRDT model is expressed as:

[0051] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects: the system and method for modeling and analyzing tail gas emissions from the incineration process of urban solid waste provided by the present invention, the system includes a multi-software coupled full-process numerical simulation model construction module under the benchmark working condition, a multi-operating working condition simulation mechanism data acquisition module, a tail gas emission model construction module based on MIMO-LRDT and a tail gas emission analysis module based on single / double factors, the method includes constructing a full-process numerical simulation model that can fit the industrial site benchmark working condition based on the multi-software coupled full-process numerical simulation model construction module under the benchmark working condition, acquiring simulation mechanism data under multiple operating conditions through the full-process data simulation model based on the multi-operating working condition simulation mechanism data acquisition module, constructing an tail gas emission model based on the MIMO-LRDT algorithm through the tail gas emission model construction module based on MIMO-LRDT, and analyzing the tail gas emission based on the single / double factors. The exhaust gas analysis module performs single / double factor analysis on the mapping relationship between the operating variables and the exhaust gas emission concentration based on the exhaust emission model. The full-process numerical simulation model is used to lay the foundation for analyzing the relationship between the "air distribution and material distribution" operating variables and the exhaust emission gas. Based on the full-process numerical simulation model, a twelve-factor four-level orthogonal experiment is designed and implemented to obtain simulation mechanism data under multiple operating conditions to provide support for the construction of a data-driven model. The MIMO-LRDT algorithm is used to establish a multi-input and multi-output data-driven model with the "air distribution and material distribution" operating variables as input and CO, CO2, O2, SO2 and NOx as output. Compared with the comparative algorithm, it has certain advantages in model fitting and model structure. Based on the established MIMO-LDRT data-driven model, the single / double factor strategy is used to analyze the relationship between the operating variables and exhaust gas emissions, providing corresponding guidance for the optimization of MSWI process operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0053] Figure 1 is a schematic diagram of the process flow of municipal solid waste incineration;

[0054] FIG2 is a schematic diagram of the structure of a system for modeling and analyzing tail gas emissions from a municipal solid waste incineration process according to an embodiment of the present invention;

[0055] FIG3 is a schematic diagram of the gas component distribution results during the combustion process on the grate;

[0056] FIG4 is a schematic diagram of the simulation results of the combustion process in the furnace;

[0057] Figure 5a is a schematic diagram of CO emissions obtained by simulation;

[0058] Figure 5b is a schematic diagram of CO2 emissions obtained by simulation;

[0059] Figure 5c is a schematic diagram of O2 emissions obtained by simulation;

[0060] Figure 5d is a schematic diagram of SO2 emissions obtained by simulation;

[0061] Figure 5e is a schematic diagram of NOx emissions obtained by simulation;

[0062] Figure 6a is a schematic diagram of feed amount analysis;

[0063] Figure 6b is a schematic diagram of the analysis of primary air temperature 2;

[0064] Figure 7a is a schematic diagram of feed rate and grate speed analysis;

[0065] Figure 7b is a schematic diagram of the analysis of feed rate and primary air temperature 1;

[0066] FIG8 is a flow chart of a method for modeling and analyzing tail gas emissions from a municipal solid waste incineration process according to an embodiment of the present invention. DETAILED DESCRIPTION

[0067] The purpose of the present invention is to provide a system and method for modeling and analyzing tail gas emissions from a municipal solid waste incineration process, which can realize modeling and analysis of tail gas emissions from a municipal solid waste incineration process.

[0068] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0069] The municipal solid waste incineration process is shown in Figure 1.

[0070] Based on the process and control status of a certain MSWI plant, the present invention adopts feed rate, primary air temperature, primary air flow, secondary air temperature and secondary air flow as key operating variables to construct a full-process numerical simulation model and a data-driven model.

[0071] As shown in Figure 2, the exhaust emission modeling and analysis system for the municipal solid waste incineration process provided by an embodiment of the present invention includes a multi-software coupled full-process numerical simulation model construction module under a reference condition, a multi-operating condition simulation mechanism data acquisition module, an exhaust emission model construction module based on MIMO-LRDT, and an exhaust emission analysis module based on a single / double factor. The multi-software coupled full-process numerical simulation model construction module under the reference condition is connected to the multi-operating condition simulation mechanism data acquisition module, the multi-operating condition simulation mechanism data acquisition module is connected to the exhaust emission model construction module based on MIMO-LRDT, and the exhaust emission model construction module based on MIMO-LRDT is connected to the exhaust emission analysis module based on a single / double factor.

[0072] A module for constructing a full-process numerical simulation model using multiple software coupling under the benchmark operating conditions: FLIC is used to simulate the solid-phase combustion process on the grate, Fluent is used to simulate the gas-phase combustion process in the furnace, and Aspen Plus is used to simulate the waste heat exchange and flue gas purification stages, thereby constructing a full-process numerical simulation model that can fit the industrial field benchmark operating conditions;

[0073] The multi-operation condition simulation mechanism data acquisition module: performs orthogonal experimental design for the "partial air distribution and material distribution" operation variables, and obtains simulation mechanism data under multiple operating conditions based on the numerical simulation model under the above-mentioned benchmark working conditions;

[0074] The exhaust emission model construction module based on MIMO-LRDT: uses the "air distribution and material distribution" operation variables as input to construct an exhaust emission model based on the MIMO-LRDT algorithm;

[0075] The exhaust emission analysis module based on single / double factors: Based on the above exhaust emission model, single / double factor analysis is performed on the mapping relationship between the operating variables and the exhaust emission concentration to support the control strategy of the industrial site.

[0076] As shown in FIG8 , the present invention further provides a method for modeling and analyzing tail gas emissions from a municipal solid waste incineration process, which is applied to the above-mentioned system for modeling and analyzing tail gas emissions from a municipal solid waste incineration process, and includes the following steps:

[0077] Step 1: Based on the multi-software coupled full-process numerical simulation model construction module under the benchmark working conditions, a full-process numerical simulation model that can fit the industrial site benchmark working conditions is constructed;

[0078] Step 2: Based on the multi-operating condition simulation mechanism data acquisition module, the simulation mechanism data under multiple operating conditions is acquired through the full-process data simulation model;

[0079] Step 3: Construct an exhaust emission model based on the MIMO-LRDT algorithm through the exhaust emission model construction module based on the MIMO-LRDT algorithm;

[0080] Step 4: Perform single / double factor analysis on the mapping relationship between the operating variables and the exhaust emission concentration based on the exhaust emission model through the exhaust emission analysis module based on single / double factors.

[0081] In step 1, a full-process numerical simulation model that can fit the industrial site benchmark working conditions is constructed based on the multi-software coupled full-process numerical simulation model construction module under the benchmark working conditions, which specifically includes the following steps:

[0082] Step 101: Solid phase combustion process on the grate based on FLIC simulation;

[0083] Step 102: Simulating the gas phase combustion process in the furnace using Fluent;

[0084] Step 103: Waste heat exchange and other flue gas purification processes based on Aspen Plus simulation.

[0085] In step 101, the solid phase combustion process on the grate based on FLIC simulation is as follows:

[0086] Under the assumption that the fuel on the grate is a homogeneous porous medium and the porosity remains unchanged during the combustion process, the flow of particulate matter in the furnace is not considered, the grate bed is considered to move forward at a constant speed, the MSW is composed of moisture, volatile matter, fixed carbon and ash, and the gas phase components only consider CO, CO2, CH4, H2, H2O, NO, HCN, NH3, N2 and O2, the solid phase combustion process on the grate is simulated using FLIC, which includes the solid phase MSW combustion model, basic conservation equations and component transport and heat radiation conversion equations.

[0087] The solid-phase MSW combustion process includes water evaporation, volatile analysis, volatile combustion and coke oxidation stages. MSW is pushed from the feeder to the grate. After passing through the high-temperature heat radiation in the furnace and the furnace wall, the water gradually evaporates at a rate of:

[0088] Where R evp is the water evaporation rate, S a is the particle surface area, h s is the convective mass transfer coefficient, C w,s and C w,g are the moisture content in MSW and mixed gas, Q cr To absorb heat during radiation and convection heat transfer, H evp The heat required for water evaporation is absorbed, T s is the MSW temperature. After the water evaporates completely, the MSW reaches the volatile analysis temperature. The release process of the main volatiles is: MSW→Volatile(C m H n,CO,CO2,H2O)+Char (2)

[0089] The volatile gas mixes with air and burns. The corresponding combustion reaction is: C m H n +(m / 2+n / 4)O2→mCO+n / 2H2O (3)CO+1 / 2O2→CO2 (4)

[0090] Its combustion rate is the minimum of the kinetic rate and the mixing rate. MSW gradually turns into coke as the volatile matter is released. It further reacts with air to produce CO and CO2, which is: C + αO2 → 2(1-α)CO + (2α-1)CO2 (5)

[0091] Wherein, α is the stoichiometric coefficient, and 0.5≤α≤1.

[0092] In step 102, the gas phase combustion process in the furnace is simulated based on Fluent, specifically:

[0093] The FLIC output is regarded as the boundary condition of the Fluent numerical simulation. In Fluent, the combustion component CmHn is replaced by CH4. The second-order upwind scheme and SIMPLE algorithm are used to discretize and solve the gas phase component combustion equation. The thermal radiation DO model is:

[0094] Where I is the radiation intensity, and are the position vector and direction vector respectively, a is the absorption coefficient, σ s is the diffusion coefficient, n is the refractive index, σ is the Boltzmann constant, Φ is the phase function, Ω′ is the fixed angle, and the standard k-ε model for turbulent gas flow is:

[0095] Where ρ is the gas density, k is the turbulent kinetic energy, and u i is the speed, x i and x j The coordinate system μ is the turbulent viscosity, and the eddy dissipation conceptual model for solving the interaction between gas flow and combustion chemical reaction is:

[0096] Where Y i is the mass fraction of substance i, is the diffusion flux of the substance, R i is the net production rate, S iTo generate additional rates, the thermal radiation distribution data after the Fluent simulation is used as the input of FLIC for iterative coupling. When the temperature difference between the two outputs meets the set error, the visualized temperature field, velocity field, and concentration field in the furnace are output.

[0097] In step 103, the waste heat exchange and flue gas purification processes simulated based on Aspen Plus are as follows:

[0098] Under the assumptions that all reactions occurring in the furnace can reach equilibrium, the furnace temperature and pressure are constant, the MSW ash is an inert component that does not participate in any reaction, the effect of MSW particle size on the combustion reaction is ignored, and pressure and gas loss and leakage are not considered, the flue gas temperature and flue gas composition simulated by FLIC, the furnace temperature simulated by Fluent, and process data from the actual industrial site (MSW ash content, secondary air flow rate, secondary air temperature, urea solution, economizer feed water, calcium hydroxide solution, activated carbon and reclaimed water) are input into Aspen Plus to simulate other stages such as MSW storage and incineration, heat exchange in the waste heat boiler, flue gas treatment and flue gas emission;

[0099] In the MSW storage and incineration process, the flue gas temperature and flue gas components (including H2O, N2, O2, H2, CO, CO2, CH4 and S) output by FLIC, the furnace temperature output by Fluent, the MSW ash content obtained based on laboratory tests, and the secondary air temperature, secondary air flow rate and urea solution dosage collected from actual industrial sites are used as inputs. The Gibbs reactor RDibbs is used to simulate the combustion process in the furnace, and the stoichiometric reactor Rstoic1 is used to simulate the denitrification reaction. The reactions involved are as follows:

[0100] In the heat exchange process stage of the waste heat boiler, two stream heat exchanger modules are used to simulate the heat exchange process between the high-temperature flue gas and the superheater and economizer to obtain the cooled furnace outlet flue gas G1. In the flue gas treatment process stage, a stoichiometric reactor is used to simulate the deacidification process of the flue gas G1. The simulation results are as follows:

[0101] The mixer module is used to simulate the adsorption process of activated carbon on pollutants such as heavy metals and dioxins in the flue gas, and the component separator module is used to simulate the bag filter. Finally, the diverter module is used to simulate the process of fly ash entering the ash silo to obtain the treated flue gas G2. In the flue gas emission stage, the compressor module is used to simulate the flue gas G3 entering the atmosphere.

[0102] In step 2, the simulation mechanism data acquisition module based on multiple operating conditions obtains simulation mechanism data under multiple operating conditions through the full-process data simulation model, specifically:

[0103] Based on the above multi-software coupling simulation strategy, combined with the process of a Beijing MSWI plant, a four-level orthogonal experimental design was conducted with 12 factors, including one non-operated variable (MSW component) and 11 operated variables (feed rate, grate speed, primary air temperature 1, primary air temperature 2, primary air temperature 3, primary air flow 1, primary air flow 2, primary air flow 3, primary air flow 4, secondary air temperature, and secondary air flow). The values ​​of each factor level are shown in Table 1.

[0104] Table 1 Experimental design of each factor level value

[0105] Based on the above experimental design, 64 sets of experimental schemes were obtained. The experiments were implemented using a multi-software coupled full-process numerical simulation model to obtain simulation mechanism data under multiple operating conditions.

[0106] In step 3, the exhaust emission model based on the MIMO-LRDT algorithm is constructed through the exhaust emission model construction module based on the MIMO-LRDT algorithm, specifically:

[0107] Based on the simulation mechanism data under the above multiple operating conditions, a multi-input and multi-output linear regression decision tree algorithm is used to establish an exhaust emission model including CO, CO2, O2, SO2 and NOx emissions. The obtained simulation mechanism data is recorded as

[0108] The mean square error of the target value is calculated by traversing the simulation mechanism data:

[0109] Where, represents the loss function value of MSE, (n,i) represents the i-th eigenvalue of the n-th sample in the k-th iteration;

[0110] Based on the calculation results, select the minimum MSE as the first non-leaf node As the splitting variable, it is:

[0111] Repeat the above process and obtain (T / 2-1) intermediate nodes based on the minimum number of samples θ set by experience

[0112] The linear regression method is used to calculate the predicted output of the CART leaf node, which is:

[0113] Where, and are the output, input and weight matrix of the tth leaf node respectively;

[0114] Obtaining the weight vector is reduced to solving a class of overdetermined matrix equations, and using the regularized least squares loss function, which is:

[0115] In the formula, λ represents the regular term coefficient, and λ≥0. The above loss function is The gradient of is expressed as:

[0116] make have to:

[0117] According to the obtained weight vector, the leaf node prediction value is calculated. Considering all leaf nodes, the MIMO-LRDT model is expressed as:

[0118] In order to explore the effects of operating variables such as feed rate, grate speed, primary air temperature 1, primary air temperature 2, primary air temperature 3, primary air flow 1, primary air flow 2, primary air flow 3, primary air flow 4, secondary air temperature and secondary air flow on CO, CO2, O2, SO2 and NOx in exhaust gas emissions, single / double factor analysis of the above operating variables was carried out based on the constructed MIMO-LRDT exhaust emission model.

[0119] The present invention provides an embodiment for simulation verification. A numerical simulation is performed based on a certain MSWI power plant. The MSW component parameters and incinerator operating parameters under the benchmark conditions are shown in Table 2 and Table 3 respectively.

[0120] Table 2 MSW component parameters

[0121] Table 3 Incinerator operating parameters under baseline conditions

[0122] Results and discussion of the multi-software coupled full-process numerical simulation model construction under benchmark conditions:

[0123] The gas component distribution of the combustion process on the grate based on FLIC simulation is shown in Figure 3. As shown in Figure 3, when the MSW gradually moves on the grate, the water evaporation rate increases with the high-temperature thermal radiation, causing the mass fraction of H2O in the flue gas to rise continuously before 2.0 m, and then briefly decrease after reaching the peak. The reason for the brief rise of H2O at 4.0 m to 5.0 m is that the primary air carries some H2O. Finally, as the combustion process proceeds, it drops to 0 near 8.0 m. The mass fractions of CO, CmHn and H2 increase with the increase of the volatile matter release rate. CO2 is produced by the mixed combustion of volatile matter, coke and O2, which causes the increase of CO2 mass fraction and the decrease of O2 mass fraction.

[0124] Figure 4 shows the temperature distribution diagram and the mass fraction cloud diagram of O2 and CO2 in the furnace during the combustion process simulated by Fluent. As shown in Figure 4, the combustion process consumes a large amount of O2, thereby generating CO2. Therefore, the temperature and CO2 mass fraction are highest in the middle of the furnace, and the O2 mass fraction is lowest. The flue gas gradually flows along the furnace wall to the furnace outlet, resulting in a gradual decrease in temperature compared to the middle of the furnace. At the same time, combustion is basically completed during the flue gas flow process, which increases the O2 mass fraction and decreases the CO2 mass fraction.

[0125] The numerical simulation results of waste heat exchange and flue gas purification processes based on Aspen Plus simulation are shown in Table 4;

[0126] Table 4 Aspen Plus numerical simulation results

[0127] It can be seen from Table 4 that with the physical and chemical reactions occurring in the flue gas treatment stage, the O2 concentration and CO2 concentration of the flue gas G3 can reach the national standard emission level.

[0128] 5a-5e. It can be seen from the incineration mechanism that CO2 is produced by the mixed combustion of volatile matter, coke and O2, and the CO2 emission concentration is inversely proportional to the O2 and CO emission concentrations. This mechanism is reflected in Figures 5a, 5b and 5c. Therefore, the effectiveness of the multi-software coupled full-process numerical simulation model under the benchmark conditions established by the present invention and the designed orthogonal experiment are verified. However, the O2 emission concentration at sampling points 14 and 16 in Figure 5c of the obtained simulation mechanism data exceeds the standard value of oxygen content in the air (21%). Therefore, the data is identified as abnormal data and needs to be eliminated in the modeling process.

[0129] Results and discussion of the exhaust emission model construction based on MIMO-LRDT: Based on the results of the multi-operating condition simulation mechanism data acquisition module described above, 62 sets of data for establishing the MIMO-LRDT model were obtained after preprocessing. In order to verify the effectiveness of the model constructed by the present invention, back propagation neural network (BPNN), decision tree (CART) and random forest (RF) were used to establish a multi-input single-output model for comparative experiments. Among them, the parameters of the MIMO-LRDT model were set as follows: the minimum number of samples is 5, the regularization coefficient is 0.5, the parameters of the CART model are set as follows: the minimum number of samples is 5, the parameters of the RF model are set as follows: the minimum number of samples is 5, the number of decision trees is 100, and the parameters of the GBDT model are set as follows: the minimum number of samples is 5, the number of iterations is 5, and the learning rate is 0.5;

[0130] The root mean square error (RMSE) indicator is used to evaluate the performance of the above model. The calculation formula is as follows:

[0131] Where, and y i Represent the model prediction value and true value respectively; N represents the number of samples;

[0132] The statistical results of RMSE of the model test set are shown in Table 5;

[0133] Table 5 RMSE statistical results of the model test set

[0134] As can be seen from Table 5, there are differences in the statistical results of the output of different model test sets. Among them, the MIMO-LRDT model has the best effect on the statistical results of NOx output and average value; the CART model has the best effect on the statistical results of CO output; the RF model has the best statistical results on the statistical results of CO2 output; the GBDT model has the best effect on the statistical results of O2 and SO2 output; the exhaust emission model established by the present invention has the smallest average error.

[0135] The proposed MIMO-LRDT model is an improvement on the CART model. It converts leaf node mean outputs into weights, improving model accuracy while also smoothing the model output, making it more suitable for controlled plant applications. Furthermore, compared to the multi-input, single-output model established using comparative methods, the proposed MIMO-LRDT model demonstrates significant advantages in statistical results, validating its effectiveness.

[0136] Results and discussion of tail gas emission analysis based on single / dual factors: For the constructed MIMO-LRDT model, other operating variables were fixed at the baseline operating conditions, and single-factor analysis was performed by varying the feed rate and primary air temperature within the set range. The curves of the variation of different gas concentrations in the tail gas are shown in Figures 6a and 6b.

[0137] As shown in Figure 6a, as the feed rate gradually increases, the emission concentrations of O2 and NOx gradually decrease, the CO emission concentration first decreases and then increases, and CO2 and SO2 show an overall upward trend. As shown in Figure 6b, as the primary air temperature 2 gradually increases, the emission concentrations of CO, CO2 and SO2 first decrease and then increase, while O2 and NOx show a trend of first increasing and then decreasing. In summary, the emission concentration change trend in the tail gas is basically consistent with the cognition of the MSWI process mechanism.

[0138] Similarly, a two-factor analysis was performed by simultaneously changing the grate speed and feed rate, as well as the primary air temperature 1 and feed rate within the set range. The CO emission concentration change curves in the tail gas are shown in Figures 7a and 7b.

[0139] As shown in Figure 7, compared with the feed rate, the grate speed and primary air temperature 1 have less impact on the CO emission concentration in the exhaust gas. Therefore, the feed rate needs to be reasonably selected to ensure that the gas emissions in the exhaust gas meet the standards and increase the economic benefits of the MSWI power plant.

[0140] The present invention provides a system and method for modeling and analyzing tail gas emissions from a municipal solid waste incineration process. The system includes a multi-software coupled full-process numerical simulation model construction module under a benchmark condition, a multi-operating condition simulation mechanism data acquisition module, a tail gas emission model construction module based on MIMO-LRDT, and a tail gas emission analysis module based on a single / double factor. The method includes constructing a full-process numerical simulation model that can fit the industrial site benchmark condition based on the multi-software coupled full-process numerical simulation model construction module under the benchmark condition, acquiring simulation mechanism data under multiple operating conditions through the full-process data simulation model based on the multi-operating condition simulation mechanism data acquisition module, constructing a tail gas emission model based on the MIMO-LRDT algorithm through the tail gas emission model construction module based on the MIMO-LRDT, and analyzing the tail gas emission model based on the tail gas emission analysis module based on the single / double factor. A single / double factor analysis was performed on the mapping relationship between the operating variables and the exhaust emission concentration. The full-process numerical simulation model was used to lay the foundation for analyzing the relationship between the "air distribution and material distribution" operating variables and the exhaust emission gases. Based on the full-process numerical simulation model, a twelve-factor four-level orthogonal experiment was designed and implemented to obtain simulation mechanism data under multiple operating conditions to provide support for the construction of a data-driven model. The MIMO-LRDT algorithm was used to establish a multi-input and multi-output data-driven model with the "air distribution and material distribution" operating variables as input and CO, CO2, O2, SO2 and NOx as output. Compared with the comparative algorithm, it has certain advantages in model fitting and model structure. Based on the established MIMO-LDRT data-driven model, a single / double factor strategy was used to analyze the relationship between the operating variables and exhaust emissions, providing corresponding guidance for the optimization of MSWI process operation.

[0141] The present invention uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only intended to help understand the method and core concept of the present invention. At the same time, those skilled in the art will find that the specific implementation methods and application scopes may vary based on the concept of the present invention. In summary, the contents of this specification should not be construed as limiting the present invention.

Claims

1. A modeling and analysis system for tail gas emissions from municipal solid waste incineration, characterized in that: include: A multi-software coupling full-process numerical simulation model construction module under benchmark conditions, a multi-operating condition simulation mechanism data acquisition module, an exhaust emission model construction module based on MIMO-LRDT, and an exhaust emission analysis module based on single / double factors, wherein the multi-software coupling full-process numerical simulation model construction module under benchmark conditions is connected to the multi-operating condition simulation mechanism data acquisition module, the multi-operating condition simulation mechanism data acquisition module is connected to the exhaust emission model construction module based on MIMO-LRDT, and the exhaust emission model construction module based on MIMO-LRDT is connected to the exhaust emission analysis module based on single / double factors; The multi-software coupled full-process numerical simulation model construction module under the benchmark working condition is used to construct a full-process numerical simulation model that can fit the industrial field benchmark working condition; The multi-operating condition simulation mechanism data acquisition module is used for the full-process data simulation model to acquire simulation mechanism data under multiple operating conditions; The exhaust emission model construction module based on MIMO-LRDT is used to construct an exhaust emission model based on the MIMO-LRDT algorithm; The exhaust emission analysis module based on single / double factors is used to perform single / double factor analysis on the mapping relationship between the operating variables and the exhaust emission concentration based on the exhaust emission model.

2. A method for modeling and analyzing tail gas emissions from the incineration process of urban solid waste, applied to the modeling and analyzing system for tail gas emissions from the incineration process of urban solid waste as claimed in claim 1, characterized in that: The steps include: Step 1: Based on the multi-software coupling full-process numerical simulation model construction module under the benchmark working conditions, a full-process numerical simulation model that can fit the industrial site benchmark working conditions is constructed; Step 2: Based on the multi-operating condition simulation mechanism data acquisition module, the simulation mechanism data under multiple operating conditions is acquired through the full-process data simulation model; Step 3: Construct an exhaust emission model based on the MIMO-LRDT algorithm through the exhaust emission model construction module based on the MIMO-LRDT; Step 4: Perform single / double factor analysis on the mapping relationship between the operating variables and the exhaust emission concentration based on the exhaust emission model through the exhaust emission analysis module based on single / double factors.

3. The method for modeling and analyzing tail gas emissions from the incineration process of municipal solid waste according to claim 2, characterized in that: In step 1, a full-process numerical simulation model that can fit the industrial site benchmark conditions is constructed based on a multi-software coupled full-process numerical simulation model construction module under benchmark conditions, which specifically includes the following steps: Step 101: Solid phase combustion process on the grate based on FLIC simulation; Step 102: gas phase combustion process in the furnace based on Fluent simulation; Step 103: Waste heat exchange and other flue gas purification processes based on Aspen Plus simulation.

4. The method for modeling and analyzing tail gas emissions from the incineration process of municipal solid waste according to claim 3, characterized in that: In step 101, the solid phase combustion process on the grate based on FLIC simulation is specifically as follows: The solid phase combustion process on the grate is simulated using FLIC, which includes the solid phase MSW combustion model, basic conservation equations, and component transport and heat radiation conversion equations; The solid phase MSW combustion process includes water evaporation, volatile analysis, volatile combustion and coke oxidation stages. MSW is pushed from the feeder to the grate. After high temperature heat radiation in the furnace and the furnace wall, the water gradually evaporates at a rate of: In the formula, R evp is the water evaporation rate, S a is the particle surface area, h s is the convective mass transfer coefficient, C w,s and C w,g are the moisture content in MSW and mixed gas, Q cr The heat absorbed during the radiation and convection heat transfer process, H evp The heat required for water evaporation, T s is the MSW temperature. After the water evaporates completely, the MSW reaches the volatilization analysis temperature. The release process of the main volatiles is: MSW→Volatile(C m H n ,CO,CO2,H2O)+Char (2) The volatile gas mixes with air and burns. The corresponding combustion reaction is: C m H n +(m / y2+n / y4)O2→mCO+n / 2H2O (3) CO+1 / 2O2→CO2 (4) Its combustion rate is the minimum of the kinetic rate and the mixing rate. MSW gradually turns into coke as the volatile matter is precipitated, and further reacts with air to generate CO and CO2, which is: C+αO2→2(1-α)CO+(2α-1)CO2 (5) Wherein, α is the stoichiometric coefficient, and 0.5≤α≤1.

5. The method for modeling and analyzing tail gas emissions from the incineration process of municipal solid waste according to claim 4, characterized in that: In step 102, the gas phase combustion process in the furnace is simulated based on Fluent, specifically: The FLIC output is regarded as the boundary condition of Fluent numerical simulation. The combustion component CmHn is replaced by CH4 in Fluent. The second-order upwind scheme and SIMPLE algorithm are used to discretize and solve the gas phase component combustion equation. The thermal radiation DO model is: Where I is the radiation intensity, and are the position vector and direction vector respectively, a is the absorption coefficient, σ s is the diffusion coefficient, n is the refractive index, σ is the Boltzmann constant, Φ is the phase function, Ω′ is the fixed angle, and the standard k-ε model for solving turbulent gas flow is: Where ρ is the gas density, k is the turbulent kinetic energy, and u i is the speed, x i and x j The coordinate system μ is the turbulent viscosity, and the eddy dissipation conceptual model for solving the interaction between gas flow and combustion chemical reaction is: Where Y i is the mass fraction of substance i, is the diffusion flux of the substance, R i is the net production rate, S i In order to generate additional rates, the thermal radiation distribution data after the Fluent simulation is used as the input of FLIC for iterative coupling. After the temperature difference between the two outputs satisfies the set error, the visualized temperature field, velocity field, and concentration field in the furnace are output.

6. The method for modeling and analyzing tail gas emissions from the incineration process of municipal solid waste according to claim 5, characterized in that: In step 103, the waste heat exchange and flue gas purification processes simulated based on Aspen Plus are as follows: The inputs are the flue gas temperature and flue gas components output by FLIC, the furnace temperature output by Fluent, the MSW ash content obtained based on the test, and the secondary air temperature, secondary air flow rate and urea solution dosage collected from the actual industrial site. The flue gas components include H2O, N2, O2, H2, CO, CO2, CH4 and S, the reactions involved in the combustion process in the furnace simulated by the Gibbs reactor RBibbs and the denitrification reaction simulated by the stoichiometric reactor Rstoic1 are: In the heat exchange process stage of the waste heat boiler, two stream heat exchanger modules are used to simulate the heat exchange process between the high-temperature flue gas and the superheater and economizer to obtain the furnace outlet flue gas G1 after cooling. In the flue gas treatment process stage, a stoichiometric reactor is used to simulate the deacidification process of the flue gas G1 as follows: The mixer module is used to simulate the adsorption process of activated carbon on pollutants such as heavy metals and dioxins in flue gas, and the component separator module is used to simulate the bag filter. Finally, the diverter module is used to simulate the process of fly ash entering the ash bin to obtain the treated flue gas G2. In the flue gas emission stage, the compressor module is used to simulate the flue gas G3 entering the atmosphere.

7. The method for modeling and analyzing tail gas emissions from the incineration process of municipal solid waste according to claim 6, characterized in that: In step 2, the simulation mechanism data acquisition module based on multiple operating conditions obtains the simulation mechanism data under multiple operating conditions through the full process data simulation model, specifically: One non-operated variable and 11 operated variables were selected for a four-level orthogonal experimental design, in which the non-operated variable was the MSW component, and the operated variables included feed rate, grate speed, primary air temperature 1, primary air temperature 2, primary air temperature 3, primary air flow 1, primary air flow 2, primary air flow 3, primary air flow 4, secondary air temperature and secondary air flow. Based on the orthogonal experimental design, the simulation mechanism data under multiple operating conditions were obtained through the full-process data simulation model.

8. The method for modeling and analyzing tail gas emissions from the incineration process of municipal solid waste according to claim 7, characterized in that: In step 3, an exhaust emission model based on the MIMO-LRDT algorithm is constructed by using an exhaust emission model construction module based on the MIMO-LRDT algorithm, specifically: Based on the simulation mechanism data under the above multiple operating conditions, a multi-input multi-output linear regression decision tree algorithm is used to establish an exhaust emission model including CO, CO2, O2, SO2 and NOx emission gases. The obtained simulation mechanism data is recorded as The simulation mechanism data is traversed to calculate the mean square error of the target value, which is: In the formula, represents the loss function value of MSE, (n,i) represents the i-th eigenvalue of the n-th sample in the k-th iteration; Based on the calculation results, select the minimum MSE as the first non-leaf node As the splitting variable, it is: Repeat the above process and obtain (T / 2-1) intermediate nodes according to the minimum number of samples θ set by experience The linear regression method is used to calculate the predicted output of the CART leaf node: In the formula, and are the output, input and weight matrix of the tth leaf node respectively; Obtaining the weight vector is classified as solving a type of overdetermined matrix equation, and the regularized least squares loss function is used, which is: In the formula, λ represents the value of the regular term coefficient, and λ≥0. The above loss function is The gradient representation for: make have to: According to the obtained weight vector, the leaf node prediction value is calculated. Considering all leaf nodes, the MIMO-LRDT model is expressed as:

Images