MSW process representation system based on virtual data and mechanism knowledge driving

Abstract

The invention provides an MSW process representation system based on virtual data and mechanism knowledge driving, and relates to the technical field of MSW processes. Comprising a full-flow numerical simulation modeling module for key controlled variables and easily-measured pollutants, a multi-software coupling full-flow numerical model under a reference operation condition is established according to the key controlled variables, conventional pollutants, physical parameters and manipulated variables; the expected quality virtual simulation mechanism data acquisition module based on multi-experiment design and implementation is used for changing manipulated variables in reference working condition values of the MSWI process at any time so as to carry out an experiment at any time to obtain virtual data under multiple working conditions. And the interpretable series-parallel hybrid integrated fuzzy forest full-process representation model module under the wide information domain is used for constructing an MSWI process full-process model according to the virtual data under multiple working conditions. According to the method, the problem that in the prior art, the information domain range represented by the real historical process data of the MSWI process is limited is solved.

Description

Technical Field

[0001] The present invention relates to the field of MSW process technology, and in particular to an MSW process characterization system driven by virtual data and mechanism knowledge. Background Art

[0002] Municipal solid waste (MSW) continues to increase at an annual rate of 8% to 10%. Recycling MSW is a key issue in developing environmentally friendly cities. MSW incineration (MSWI) is a crucial component of renewable energy recycling in cities worldwide and has become an effective solution to the "garbage siege" phenomenon.

[0003] Because the actual operation of the MSWI process is accompanied by high-temperature, high-pressure equipment, toxic and hazardous gases, etc., it is difficult to construct its physical model in the laboratory. This makes the construction of its numerical simulation model one of the effective solutions for simulating the multi-operating characteristics of the actual process. The solid-phase combustion process of the MSWI process mainly adopts the bed combustion model (FLIC) and discrete element model (DEM), the gas-phase combustion process mainly adopts Fluent software, and the multi-process stage simulation mainly adopts AspenPlus software. Due to the complementarity between the above numerical simulation software, the use of coupling strategies has become an effective means to improve the realism of the simulation, such as the coupling of Fluent and Aspen Plus, Fluent and Hydrocarbon Cracking Simulation Software (COILSM), FLIC and Fluent; however, the above coupling is mainly used to simulate solid-phase and gas-phase combustion.

[0004] The complex dynamic characteristics and ambiguous internal mechanisms of real-world nonlinear systems make it difficult to develop a representational model that can replace the original system through purely theoretical means. Therefore, academics often transform industrial process modeling problems into data-driven models tailored to specific control requirements, using manipulated variables as input. For the MSWI process, single-input, single-output (SISO) process models include TS fuzzy neural networks and least squares-support vector regression for furnace temperature, Mamdani fuzzy models for oxygen content, and radial basis function neural networks for steam flow. Multi-input, multi-output (MIMO) process models include ARX models for flue gas oxygen content and steam flow, and TS fuzzy neural networks for furnace temperature, flue gas oxygen content, and steam flow. Furthermore, full-process models encompass key controlled variables and easily measurable pollutants. However, these models are primarily based on single-operating-condition real-world industrial data with a narrow information domain and fail to reflect the multi-operating-condition characteristics of dynamic MSWI processes. Research on combining high-reliability virtual simulation mechanism data with a wide information domain to construct a full-process representational model driven by virtual and real data has yet to be reported. Summary of the Invention

[0005] In order to overcome the shortcomings of the existing technology, the purpose of the present invention is to provide an MSW process characterization system driven by virtual data and mechanism knowledge. The present invention solves the problem of limited information domain range that can be represented by real historical process data of MSW process in the existing technology.

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

[0007] A MSW process characterization system based on virtual data and mechanism knowledge, including:

[0008] The modules are a full-process numerical simulation modeling module for key controlled variables and easily measured pollutants, a data acquisition module for the expected quality virtual simulation mechanism based on multiple experimental design and implementation, and an interpretable series-parallel hybrid integrated fuzzy forest full-process characterization model module under a wide information domain.

[0009] The full-process numerical simulation modeling module for key controlled variables and easily measured pollutants is used to establish a multi-software coupled full-process numerical model under benchmark operating conditions based on key controlled variables, conventional pollutants, physical parameters and manipulated variables. As shown below,

[0010]

[0011] in, It represents the manipulated variable input vector under multiple working conditions, which consists of feed rate, grate speed, primary air volume, secondary air volume, boiler feed water volume, urea, slaked lime, activated carbon, etc. represents the physical parameter input vector consisting of the incinerator size, grate length and width, etc.; and They represent the outputs of the furnace temperature, boiler steam flow, G1 oxygen content, and G3 oxygen content models under multiple operating conditions; Table 1 shows the output vector of G1 pollutants under multiple working conditions. and Respectively represent the output values ​​of NOx, SO2, HCl, CO, CO2 and particulate matter (PM) at G1 under multiple working conditions; represents the output vector of the G3 pollutant model under multiple operating conditions, and Respectively represent the output values ​​of NOx, SO2, HCl, CO, CO2 and PM at G1 under multiple working conditions;

[0012] The multi-software coupled full-process numerical model under the benchmark operating conditions is used to provide the benchmark operating condition values ​​of the MSWI process. The expected quality virtual simulation mechanism data acquisition module based on multiple experimental design and implementation is used to change the manipulated variables in the benchmark operating condition values ​​of the MSWI process any number of times to conduct any number of experiments to obtain virtual data under multiple operating conditions. The interpretable series-parallel hybrid integrated fuzzy forest full-process characterization model module under a wide information domain is used to construct the MSWI process full-process model based on the virtual data under the multiple operating conditions.

[0013] Preferably, the full-process numerical simulation modeling module for key controlled variables and easily measured pollutants includes: a first model building submodule, a second model building submodule, and a third model building submodule;

[0014] The first model construction submodule is used to construct simulation models of the feeding and solid-phase combustion stages based on key controlled variables and conventional pollutants, and the second model construction submodule is used to construct a simulation model of the gas-phase combustion stage based on key controlled variables and conventional pollutants; the solid-phase combustion gas temperature and component data simulated by the first model construction submodule are transmitted to the gas-phase combustion simulation model established by the second model construction submodule, and further, the second model construction submodule transmits the gas-phase combustion temperature and other data to the solid-phase combustion stage simulation model established by the first model construction submodule, and iterates repeatedly until convergence; after convergence, the third model construction submodule is used to construct simulation models of the heat exchange, flue gas purification and exhaust emission stages based on physical parameters and manipulated variables, receive data such as temperature and gas composition obtained by the second model construction submodule, and then obtain full-process numerical simulation modeling for key controlled variables and easily measurable pollutants under benchmark operating conditions.

[0015] Preferably, the desired quality virtual simulation mechanism data acquisition module based on multiple experimental design and implementation includes:

[0016] The multiple experimental design submodule, case evaluation submodule, model simulation submodule and multi-index comprehensive evaluation submodule are connected in sequence;

[0017] The multiple experimental design submodule is used to design experiments based on the determined manipulated variables to obtain experimental cases. The case evaluation submodule is used to evaluate and judge the rationality and number of the experimental cases to obtain a first judgment result. If the first judgment result is failed, the manipulated variables are re-determined and a new experimental case is generated. If the first judgment result is passed, the experimental case is input into the model simulation submodule. The model simulation submodule is used to perform experimental instances based on the current experimental case to obtain simulation mechanism data. The multi-index comprehensive evaluation submodule is used to evaluate and judge the simulation mechanism data to obtain a second judgment result. If the second judgment result is passed, virtual data under multiple working conditions are obtained. If the second judgment result is failed, return to the multiple experimental design submodule to continue the experiment.

[0018] Preferably, the wide information domain can interpret the series-parallel hybrid integrated fuzzy forest full process representation model The expression is:

[0019]

[0020] in, It represents the manipulated variable input vector under multiple working conditions, which consists of feed rate, grate speed, primary air volume, secondary air volume, boiler feed water volume, urea, slaked lime, activated carbon, etc. represents the matrix of non-manipulated variables under multiple working conditions, which consists of MSW particle size, moisture content, etc.; and They represent the outputs of the furnace temperature, boiler steam flow, G1 oxygen content, and G3 oxygen content models under multiple operating conditions; Table 1 shows the output vector of G1 pollutants under multiple working conditions. and Respectively represent the output values ​​of NOx, SO2, HCl, CO, CO2 and particulate matter (PM) at G1 under multiple working conditions; represents the output vector of the G3 pollutant model under multiple operating conditions, and Respectively represent the output values ​​of NOx, SO2, HCl, CO, CO2 and PM at G1 under multiple working conditions.

[0021] The present invention discloses the following technical effects:

[0022] The present invention provides an MSW process characterization system driven by virtual data and mechanism knowledge, comprising: a full-process numerical simulation modeling module for key controlled variables and easily measurable pollutants, an expected quality virtual simulation mechanism data acquisition module based on multiple experimental design and implementation, and an interpretable series-parallel hybrid integrated fuzzy forest full-process characterization model module connected in sequence; the full-process numerical simulation modeling module for key controlled variables and easily measurable pollutants is used to establish a multi-software coupled full-process numerical model under a benchmark operating condition based on key controlled variables, conventional pollutants, physical parameters and manipulated variables, wherein the multi-software coupled full-process numerical model under the benchmark operating condition is used to provide the benchmark operating condition values ​​of the MSWI process, the expected quality virtual simulation mechanism data acquisition module based on multiple experimental design and implementation is used to change the manipulated variables in the benchmark operating condition values ​​of the MSWI process any number of times to conduct any number of experiments to obtain virtual data under multiple conditions, and the interpretable series-parallel hybrid integrated fuzzy forest full-process characterization model module under a wide information domain is used to construct a full-process model of the MSWI process based on the virtual data under the multiple conditions. This invention utilizes a virtual simulation mechanism data acquisition module based on multiple experimental designs. The system enables unlimited adjustments and experiments with manipulated variables in the MSWI process, generating virtual data under varying operating conditions. This approach effectively overcomes the information domain limitations inherent in relying on historical process data, encompassing a wider range of operating conditions and improving the model's adaptability to diverse operating conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] 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.

[0024] Figure 1 A schematic diagram of the structure of a MSW process characterization system based on virtual data and mechanism knowledge drive provided by an embodiment of the present invention;

[0025] Figure 2 A block diagram of a full-process numerical simulation modeling module for key controlled variables and easily measured pollutants provided by an embodiment of the present invention;

[0026] Figure 3 A block diagram of a data acquisition module for a virtual simulation mechanism of expected quality based on multiple experimental designs and implementations provided in an embodiment of the present invention;

[0027] Figure 4 A block diagram of a full-process characterization model module for interpretable serial-parallel hybrid integration fuzzy forests in a wide information domain provided by an embodiment of the present invention;

[0028] Figure 5 This is a strategy diagram of the TSFFR algorithm provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0029] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0030] 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.

[0031] like Figure 1 As shown, the present invention provides a MSW process characterization system driven by virtual data and mechanism knowledge, including:

[0032] The modules are a full-process numerical simulation modeling module for key controlled variables and easily measured pollutants, a data acquisition module for the expected quality virtual simulation mechanism based on multiple experimental design and implementation, and an interpretable series-parallel hybrid integrated fuzzy forest full-process characterization model module under a wide information domain.

[0033] The full-process numerical simulation modeling module for key controlled variables and easily measured pollutants is used to establish a multi-software coupled full-process numerical model under benchmark operating conditions based on key controlled variables, conventional pollutants, physical parameters and manipulated variables. As shown below,

[0034]

[0035] in, It represents the manipulated variable input vector under multiple working conditions, which consists of feed rate, grate speed, primary air volume, secondary air volume, boiler feed water volume, urea, slaked lime, activated carbon, etc. represents the physical parameter input vector consisting of the incinerator size, grate length and width, etc.; and They represent the outputs of the furnace temperature, boiler steam flow, G1 oxygen content, and G3 oxygen content models under multiple operating conditions; Table 1 shows the output vector of G1 pollutants under multiple working conditions. and Respectively represent the output values ​​of NOx, SO2, HCl, CO, CO2 and particulate matter (PM) at G1 under multiple working conditions; represents the output vector of the G3 pollutant model under multiple operating conditions, and Respectively represent the output values ​​of NOx, SO2, HCl, CO, CO2 and PM at G1 under multiple working conditions;

[0036] The multi-software coupled full-process numerical model under the benchmark operating conditions is used to provide the benchmark operating condition values ​​of the MSWI process. The expected quality virtual simulation mechanism data acquisition module based on multiple experimental design and implementation is used to change the manipulated variables in the benchmark operating condition values ​​of the MSWI process any number of times to conduct any number of experiments to obtain virtual data under multiple operating conditions. The interpretable series-parallel hybrid integrated fuzzy forest full-process characterization model module under a wide information domain is used to construct the MSWI process full-process model based on the virtual data under the multiple operating conditions.

[0037] Specifically, the full-process numerical simulation modeling module for key controlled variables and easily measurable pollutants provides a numerical simulation model of the MSWI process under the benchmark working conditions for the expected quality virtual simulation mechanism data acquisition module based on multiple experimental design and implementation. The expected quality virtual simulation mechanism data acquisition module based on multiple experimental design and implementation provides virtual (mechanism) data under multiple working conditions for the interpretable series-parallel hybrid integrated fuzzy forest full-process characterization model under a wide information domain, and finally establishes a full-process model of the MSWI process.

[0038] Full-process numerical simulation modeling module for key controlled variables and easily measurable pollutants: Targeting key controlled variables of the MSWI process, such as furnace temperature, flue gas oxygen content, and main steam flow, as well as easily measurable conventional pollutants such as NOx, CO, SO2, HCl, HF, and particulate matter (PM), a multi-software coupled full-process numerical model under benchmark operating conditions is established using the actual physical parameters of the MSWI plant, such as length, width, and height, and manipulated variables, such as feed rate, grate speed, primary air flow, and secondary air flow.

[0039] Expected quality virtual simulation mechanism data acquisition module based on multiple experimental design and implementation: For the full-process numerical simulation model of the benchmark working condition constructed above, multiple experimental designs and implementations are carried out, that is, a new operating condition is obtained by changing the value of any manipulated variable; and feedback, iteration, adaptation and other measures are introduced in the experimental design and result stages to ensure data quality.

[0040] Interpretable series-parallel hybrid integrated fuzzy forest full-process characterization model module under wide information domain: In order to overcome the problems of large training data requirements, complex calculations, difficult adjustment of many hyperparameters and poor interpretability of deep neural networks, based on the multi-working condition virtual simulation mechanism data obtained by the above module, a series-parallel hybrid integrated fuzzy forest model for key controlled variables and easily measured pollutants is constructed.

[0041] More specifically, this embodiment also discloses a process flow for the incineration of municipal solid waste:

[0042] 1) Solid waste fermentation stage: Sanitation vehicles transport MSW from various collection points in the city to the MSWI plant. After weighing and recording, it is dumped from the unloading platform into the unfermented area of ​​the solid waste storage tank. Then, the solid waste is mixed and stirred by the solid waste grabber and then taken to the fermentation area. It is fermented and dehydrated for 3 to 7 days to ensure the low calorific value of the MSW for incineration.

[0043] 2) Solid Waste Combustion Stage: A solid waste grabber deposits fermented MSW into the feed hopper, which pushes it into the incinerator via the feeder. After drying, combustion 1, combustion 2, and the burnout grate, the combustible components in the MSW are completely burned. The required combustion air is injected from below the grate and into the center of the furnace by primary and secondary fans. The resulting ash falls from the end of the burnout grate into the slag scoop, where it is water-cooled and then sent to the slag pool. This stage requires strict control of the flue gas temperature above 850°C, a residence time of more than 2 seconds in the furnace, and sufficient flue gas turbulence.

[0044] 3) Waste heat exchange stage: The high-temperature flue gas generated in the furnace is sucked into the waste heat boiler system by the induced draft fan, passes through the superheater, evaporator and economizer equipment, and generates high-temperature steam after heat exchange with the liquid water in the boiler drum, thereby achieving cooling treatment, so that the flue gas temperature at the waste heat boiler outlet is lower than 200℃ (i.e. flue gas G1).

[0045] 4) Steam power generation stage: The high-temperature steam generated by the waste heat boiler is used to drive the steam turbine generator, converting mechanical energy into electrical energy, achieving self-sufficiency in plant-level electricity consumption and supplying surplus electricity to the grid, thereby obtaining economic benefits.

[0046] 5) Flue gas purification stage: The flue gas purification in the MSWI process mainly includes a series of processes such as denitrification, desulfurization, heavy metal removal, dioxin adsorption and dust (particulate matter) removal, so that the generated flue gas pollutant emissions meet national standards.

[0047] 6) Flue gas emission stage: The incineration flue gas (i.e. flue gas G2) after cooling and purification is sucked by the induced draft fan and then discharged into the atmosphere through the chimney.

[0048] Further, such as Figure 2As shown, the full-process numerical simulation modeling module for key controlled variables and easily measured pollutants includes: a first model building submodule, a second model building submodule, and a third model building submodule;

[0049] The first model construction submodule is used to construct simulation models of the feeding and solid-phase combustion stages based on key controlled variables and conventional pollutants, and the second model construction submodule is used to construct a simulation model of the gas-phase combustion stage based on key controlled variables and conventional pollutants; the solid-phase combustion gas temperature and component data simulated by the first model construction submodule are transmitted to the gas-phase combustion simulation model established by the second model construction submodule, and further, the second model construction submodule transmits the gas-phase combustion temperature and other data to the solid-phase combustion stage simulation model established by the first model construction submodule, and iterates repeatedly until convergence; after convergence, the third model construction submodule is used to construct simulation models of the heat exchange, flue gas purification and exhaust emission stages based on physical parameters and manipulated variables, receive data such as temperature and gas composition obtained by the second model construction submodule, and then obtain full-process numerical simulation modeling for key controlled variables and easily measurable pollutants under benchmark operating conditions.

[0050] Specifically, first, based on the solid waste testing data from actual industrial sites and the manipulated variables simplified by expert decision-making under benchmark operating conditions, the key manipulated variables and their value ranges as inputs to the full-process model are determined;

[0051] The MSWI system was then divided into five stages: feeding, combustion, heat exchange, flue gas purification, and tail gas emission. FLIC software was used to construct simulation models for the feeding and solid-phase combustion stages, focusing on simulating the release of various flue gas components from solid waste. Fluent software was used to construct a simulation model for the gas-phase combustion stage, focusing on the pollutant conversion mechanism and key controlled variable characteristics. Aspen software was used to construct simulation models for the heat exchange, flue gas purification, and tail gas emission stages, focusing on the pollutant emission concentrations at G1 and G3, and establishing a pollutant adsorption model.

[0052] Finally, the models built by the above three software were coupled to realize full-process numerical simulation modeling. The strategy was to transfer data such as gas temperature and composition of FLIC simulated solid-phase combustion to the gas-phase combustion simulation model established by Fluent software. Furthermore, Fluent transferred data such as gas-phase combustion temperature to the solid-phase combustion stage simulation model established by FLIC software, and iterated repeatedly until convergence; after convergence, the temperature and gas composition data obtained by Fluent were transferred to the heat exchange stage model of Aspen software to complete the simulation of subsequent process stages.

[0053] Furthermore, the desired quality virtual simulation mechanism data acquisition module based on multiple experimental design and implementation includes:

[0054] The multiple experimental design submodule, case evaluation submodule, model simulation submodule and multi-index comprehensive evaluation submodule are connected in sequence;

[0055] The multiple experimental design submodule is used to design experiments based on the determined manipulated variables to obtain experimental cases. The case evaluation submodule is used to evaluate and judge the rationality and number of the experimental cases to obtain a first judgment result. If the first judgment result is failed, the manipulated variables are re-determined and a new experimental case is generated. If the first judgment result is passed, the experimental case is input into the model simulation submodule. The model simulation submodule is used to perform experimental instances based on the current experimental case to obtain simulation mechanism data. The multi-index comprehensive evaluation submodule is used to evaluate and judge the simulation mechanism data to obtain a second judgment result. If the second judgment result is passed, virtual data under multiple working conditions are obtained. If the second judgment result is failed, return to the multiple experimental design submodule to continue the experiment.

[0056] Specifically, such as Figure 3 As shown in the figure, assuming that n experimental designs are required, the design process of the first experiment is as follows: first, the manipulated variables and their level values ​​of the multiple experimental design submodule are determined based on expert experience, and experimental cases are obtained through experimental design; then, these data enter the case evaluation submodule to evaluate and judge the rationality and number of the designed cases. If they pass the test, they enter the model simulation submodule to implement the experiment to obtain simulation mechanism data; if they fail the test, they are fed back to the expert experience, and a new set of manipulated variables and their level values ​​are re-determined for experimental design, and re-evaluated and judged until qualified simulation mechanism data of the first experiment are obtained.

[0057] Correspondingly, when conducting the second experimental design, the manipulated variables and their level values ​​are adaptively adjusted based on the virtual simulation mechanism data obtained in the first experimental design, and the subsequent process is the same as that of the first experiment.

[0058] After the nth experimental design also meets the requirements, the multi-index comprehensive evaluation submodule is entered to evaluate the multi-working condition mechanism data. At this time, if the test is qualified, the virtual simulation mechanism data of the expected quality is obtained; if the test fails, it is still necessary to return to the multiple experimental design submodule.

[0059] Furthermore, the wide information domain can explain the series-parallel hybrid integrated fuzzy forest full process representation model The expression is:

[0060]

[0061] in, It represents the manipulated variable input vector under multiple working conditions, which consists of feed rate, grate speed, primary air volume, secondary air volume, boiler feed water volume, urea, slaked lime, activated carbon, etc. represents the matrix of non-manipulated variables under multiple working conditions, which consists of MSW particle size, moisture content, etc.; and They represent the outputs of the furnace temperature, boiler steam flow, G1 oxygen content, and G3 oxygen content models under multiple operating conditions; Table 1 shows the output vector of G1 pollutants under multiple working conditions. and Respectively represent the output values ​​of NOx, SO2, HCl, CO, CO2 and particulate matter (PM) at G1 under multiple working conditions; represents the output vector of the G3 pollutant model under multiple operating conditions, and Respectively represent the output values ​​of NOx, SO2, HCl, CO, CO2 and PM at G1 under multiple working conditions.

[0062] Specifically, such as Figure 4 As shown in the figure, the series-parallel hybrid integrated fuzzy forest full process characterization model module can be interpreted under the wide information domain, including the chamber temperature model, boiler steam flow model, G1 flue gas oxygen content model, G1 flue gas pollutant model, G3 flue gas oxygen content model and G3 flue gas pollutant model. These key variable models and pollutant models constitute the full process model based on the series-parallel method. Based on the virtual simulation mechanism data of expected quality obtained above, the above model is established by using the TS fuzzy forest regression (TSFFR) algorithm to enhance the interpretability of the model, including training subset division, TS decision tree (TSDT) sub-model construction and TSFFR output fusion module, such as Figure 5 shown.

[0063] Training subset partitioning submodule for virtual simulation mechanism data of expected quality Random sampling is performed to obtain The training subset of features As follows:

[0064]

[0065] in, represents the jth training subset, represents the input variable matrix, represents the output variable matrix.

[0066] Figure 5In the TSDT sub-model construction sub-module, there are a screening layer and a fuzzy inference layer. The screening layer filters and splits the training subset. To get a clear set Then determine the The input training set of TS fuzzy inference with leaf nodes Expressed as:

[0067]

[0068] in, represents the training set of fuzzy reasoning, Represents a clear set No. The input variables of leaf nodes, Indicates the The number of samples in a leaf node.

[0069] The fuzzy reasoning layer obtains the relationship between input variables and output variables through fuzzy reasoning. rule The kth IF-THEN fuzzy rule describes the local linear relationship. rule The fuzzy rules are as follows:

[0070]

[0071] in, yes No. The amount of input in a leaf, represent The membership function of and Both are input x t The fuzzy set of is the kth rule The output of the fuzzy rule, function It can be expressed by the following formula:

[0072]

[0073] Among them, ω t Represents x t The weight of .

[0074] Prediction output It can be expressed as:

[0075]

[0076] in, So x j is the predicted output for the input, is the kth part of the antecedent rule A weight.

[0077] Figure 5 In the TSFFR output fusion submodule, J TSDT sub-models and J output prediction values ​​are obtained. (denoted as A), the output is:

[0078]

[0079] in, Represents weight.

[0080] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0081] This document 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 MSW process characterization system based on virtual data and mechanism knowledge, characterized by: include: The modules are a full-process numerical simulation modeling module for key controlled variables and easily measured pollutants, a data acquisition module for the expected quality virtual simulation mechanism based on multiple experimental design and implementation, and an interpretable series-parallel hybrid integrated fuzzy forest full-process characterization model module under a wide information domain. The full-process numerical simulation modeling module for key controlled variables and easily measured pollutants is used to establish a multi-software coupled full-process numerical model under benchmark operating conditions based on key controlled variables, conventional pollutants, physical parameters and manipulated variables. As shown below, in, It represents the manipulated variable input vector under multiple working conditions, which consists of feed rate, grate speed, primary air volume, secondary air volume, boiler feed water volume, urea, slaked lime, activated carbon, etc. represents the physical parameter input vector consisting of the incinerator size, grate length and width, etc.; and They represent the outputs of the furnace temperature, boiler steam flow, G1 oxygen content, and G3 oxygen content models under multiple operating conditions; Table 1 shows the output vector of G1 pollutants under multiple working conditions. and Respectively represent the output values ​​of NOx, SO2, HCl, CO, CO2 and particulate matter (PM) at G1 under multiple working conditions; represents the output vector of the G3 pollutant model under multiple operating conditions, and Respectively represent the output values ​​of NOx, SO2, HCl, CO, CO2 and PM at G1 under multiple working conditions; The multi-software coupled full-process numerical model under the benchmark operating conditions is used to provide the benchmark operating condition values ​​of the MSWI process. The expected quality virtual simulation mechanism data acquisition module based on multiple experimental design and implementation is used to change the manipulated variables in the benchmark operating condition values ​​of the MSWI process any number of times to conduct any number of experiments to obtain virtual data under multiple operating conditions. The interpretable series-parallel hybrid integrated fuzzy forest full-process characterization model module under a wide information domain is used to construct the MSWI process full-process model based on the virtual data under the multiple operating conditions.

2. The MSW process characterization system based on virtual data and mechanism knowledge drive according to claim 1 is characterized in that: The full-process numerical simulation modeling module for key controlled variables and easily measured pollutants includes: a first model building submodule, a second model building submodule, and a third model building submodule; The first model construction submodule is used to construct simulation models of the feeding and solid-phase combustion stages based on key controlled variables and conventional pollutants, and the second model construction submodule is used to construct a simulation model of the gas-phase combustion stage based on key controlled variables and conventional pollutants; the solid-phase combustion gas temperature and component data simulated by the first model construction submodule are transmitted to the gas-phase combustion simulation model established by the second model construction submodule, and further, the second model construction submodule transmits the gas-phase combustion temperature and other data to the solid-phase combustion stage simulation model established by the first model construction submodule, and iterates repeatedly until convergence; after convergence, the third model construction submodule is used to construct simulation models of the heat exchange, flue gas purification and exhaust emission stages based on physical parameters and manipulated variables, receive data such as temperature and gas composition obtained by the second model construction submodule, and then obtain full-process numerical simulation modeling for key controlled variables and easily measurable pollutants under benchmark operating conditions.

3. The MSW process characterization system based on virtual data and mechanism knowledge drive according to claim 1 is characterized in that: The expected quality virtual simulation mechanism data acquisition module based on multiple experimental design and implementation includes: The multiple experimental design submodule, case evaluation submodule, model simulation submodule and multi-index comprehensive evaluation submodule are connected in sequence; The multiple experimental design submodule is used to design experiments based on the determined manipulated variables to obtain experimental cases. The case evaluation submodule is used to evaluate and judge the rationality and number of the experimental cases to obtain a first judgment result. If the first judgment result is failed, the manipulated variables are re-determined and a new experimental case is generated. If the first judgment result is passed, the experimental case is input into the model simulation submodule. The model simulation submodule is used to perform experimental instances based on the current experimental case to obtain simulation mechanism data. The multi-index comprehensive evaluation submodule is used to evaluate and judge the simulation mechanism data to obtain a second judgment result. If the second judgment result is passed, virtual data under multiple working conditions are obtained. If the second judgment result is failed, return to the multiple experimental design submodule to continue the experiment.

4. The MSW process characterization system based on virtual data and mechanism knowledge drive according to claim 1 is characterized in that: The interpretable series-parallel hybrid integrated fuzzy forest full-process representation model under the wide information domain The expression is: in, It represents the manipulated variable input vector under multiple working conditions, which consists of feed rate, grate speed, primary air volume, secondary air volume, boiler feed water volume, urea, slaked lime, activated carbon, etc. represents the matrix of non-manipulated variables under multiple working conditions, which consists of MSW particle size, moisture content, etc.; and They represent the outputs of the furnace temperature, boiler steam flow, G1 oxygen content, and G3 oxygen content models under multiple operating conditions; Table 1 shows the output vector of G1 pollutants under multiple working conditions. and Respectively represent the output values ​​of NOx, SO2, HCl, CO, CO2 and particulate matter (PM) at G1 under multiple working conditions; represents the output vector of the G3 pollutant model under multiple operating conditions, and Respectively represent the output values ​​of NOx, SO2, HCl, CO, CO2 and PM at G1 under multiple working conditions.

Images